The Astrophysical Journal, 586:268–285, 2003 March 20
# 2003. The American Astronomical Society. All rights reserved. Printed in U.S.A.
AN ULTRA–HIGH-RESOLUTION SURVEY OF THE INTERSTELLAR 7Li/6Li ISOTOPE RATIO
IN THE SOLAR NEIGHBORHOOD
David C. Knauth,1,2,3 S. R. Federman,1,3 and David L. Lambert4
Received 2002 September 17; accepted 2002 November 21
ABSTRACT
In an effort to probe the extent of variations in the interstellar 7Li/6Li ratio seen previously, ultra–highresolution (R 360; 000), high signal-to-noise spectra of stars in the Perseus OB2 and Scorpius OB2
associations were obtained. These measurements confirm our earlier findings of an interstellar 7Li/6Li ratio
of about 2 toward o Per, the value predicted from models of Galactic cosmic-ray spallation reactions.
Observations of other nearby stars yield limits consistent with the isotopic ratio of 12 seen in carbonaceous
chondrite meteorites. If this ratio originally represented the gas toward o Per, then to decrease the original
isotope ratio to its current value an order of magnitude increase in the Li abundance is expected, but it is not
seen. The elemental K/Li ratio is not unusual, although Li and K are formed via different nucleosynthetic
pathways. Several proposals to account for the low 7Li/6Li ratio were considered, but none seems satisfactory. Analysis of the Li and K abundances from our survey highlighted two sight lines where depletion effects
are prevalent. There is evidence for enhanced depletion toward X Per, since both abundances are lower by a
factor of 4 when compared to other sight lines. Moreover, a smaller Li/H abundance is observed toward 20
Aql, but the K/H abundance is normal, suggesting enhanced Li depletion (relative to K) in this direction.
Our results suggest that the 7Li/6Li ratio has not changed significantly during the last 4.5 billion years and
that a ratio of 12 represents most gas in the solar neighborhood. In addition, there appears to be a constant
stellar contribution of 7Li, indicating that one or two processes dominate its production in the Galaxy.
Subject headings: ISM: abundances — ISM: atoms —
open clusters and associations: individual (Perseus OB2, Scorpius OB2) —
solar neighborhood —
stars: individual (20 Aquilae, Ophiuchi, Ophiuchi, o Persei, X Persei)
duced in and then accelerated by the explosion or are
sputtered off interstellar grains and then accelerated.
A low-energy cosmic-ray (LECR) component, accelerated C and O nuclei, may also be present (Ramaty, Kozlovsky, & Lingenfelter 1996). Lithium, beryllium, and boron
are produced through the in-flight fragmentation of C and
O nuclei during collisions with ambient interstellar H and
He (Cassé, Lehoucq, & Vangioni-Flam 1995; Higdon,
Lingenfelter, & Ramaty 1998; Lemoine, Vangioni-Flam, &
Cassé 1998; Lingenfelter, Ramaty, & Kozlovsky 1998;
Vangion-Flam et al. 1998; Parizot & Drury 1999; Ramaty &
Lingenfelter 1999; Vangioni-Flam, Cassé, & Audouze
2000). The inverse of standard GCR reactions, this LECR
component originates from core collapse supernovae (Type
II supernovae [SN IIs]) grouped in a superbubble (Higdon
et al. 1998; Lingenfelter et al. 1998). Both standard GCR
and LECR/superbubble models are able to account for the
present-day abundance of 6Li, 9Be, and 10B but only 10%–
25% of 7Li and about 50% of 11B (e.g., MAR; Ramaty et al.
1997).
The rapid rise in the 7Li abundance for stars with
½Fe=H 0:5 indicates the existence of a stellar source
of 7Li. Proposed stellar sources include asymptotic giant
branch (AGB; Smith & Lambert 1989, 1990; Plez, Smith, &
Lambert 1993) and red giant branch (RGB; Smith et al.
1995) stars, with contributions from SN IIs and possibly
novae. In AGB and RGB stars, rapid transport of 7Be
(Cameron & Fowler 1971; Boothroyd, Sackmann, &
Wasserburg 1994, 1995; Wasserburg, Boothroyd, Sackmann
1995) between the He burning shell and the envelope yields
7Li. In SN IIs, the flux of neutrinos becomes so great that the
spallation of heavy nuclei, C, N, and O, into light elements
1. INTRODUCTION
One of the primary goals in astronomy centers on the origin of the elements, but a complete picture remains elusive.
Our focus is an improved understanding of light-element
synthesis through observations of interstellar lithium in diffuse clouds. Since models of big bang nucleosynthesis
(BBN) yield 10% of the current abundance of 7Li and negligible amounts of 6Li (Suzuki, Yoshi, & Beers 2000), much
theoretical effort has gone into finding the source for Li. In
the 1970s, the work of Reeves and collaborators (Reeves,
Fowler, & Hoyle 1970; Meneguzzi, Audouze, & Reeves
1971, hereafter MAR; Reeves et al. 1973; Reeves 1974) provided an alternate method of light-element production,
which plays an important role throughout the history of the
Galaxy. This process is now known as standard Galactic
cosmic-ray (GCR) spallation. Light elements are created
when interstellar C, N, and O nuclei are broken apart by
Galactic cosmic rays, the spallation process, and by -
fusion reactions. The exact nature of the cosmic-ray source
is still under debate (e.g., Ramaty et al. 2000a; Fields et al.
2001). Two commonly accepted creation mechanisms for
cosmic rays involve supernovae; p and particles are pro1 Department of Physics and Astronomy, University of Toledo, 2801
West Bancroft, Toledo, OH 43606; [email protected].
2 Current address: FUSE Science Center, Department of Physics and
Astronomy, The Johns Hopkins University, 3400 North Charles Street,
Baltimore, MD 21218; [email protected].
3 Guest Observer, McDonald Observatory, University of Texas at
Austin.
4 Department of Astronomy, University of Texas at Austin, Austin,
TX 78712; [email protected].
268
INTERSTELLAR 7Li/6Li RATIOS
can occur. Through spallation in the He and C shells, SN IIs
can produce observable amounts of both 7Li and 11B
(Woosley et al. 1990; Woosley & Weaver 1995). Another
possible thermonuclear source of 7Li is novae. Novae are
predicted to play a significant part in the 7Li abundance
(Starrfield et al. 1978; Romano et al. 2001). The nature of the
processes involved in the production of 7Li is still unclear. A
goal of our survey of the interstellar 7Li/6Li ratio is to shed
light on the importance of the various processes.
Interstellar Li was first detected in the early 1970s (Traub
& Carleton 1973). Since then there have been relatively few
reported observations of interstellar Li i. The early detections (Traub & Carleton 1973; Vanden Bout et al. 1978;
Snell & Vanden Bout 1981; Hobbs 1984; White 1986)
showed weak interstellar Li i absorption, barely discernible
from the noise, with equivalent widths (W ) on the order of
at most a few mÅ. These early observations showed that the
interstellar Li i abundance was similar along different lines
of sight. However, large uncertainties remained because of
low signal-to-noise ratios and uncertain corrections for ionization and depletion onto grains. With the advent of modern detectors and more sophisticated observing techniques,
reliable observations of 7Li and the much weaker 6Li line
were possible (Ferlet & Dennefeld 1984; Lemoine et al.
1993; Lemoine, Ferlet, & Vidal-Madjar 1995; Meyer,
Hawkins, & Wright 1993; Knauth et al. 2000; Howarth et
al. 2002).
The Li isotope ratio is crucial for studies of Galactic
chemical evolution (Reeves 1993; Steigman 1993) since it is
free from uncertain corrections (e.g., ionization and depletion). Published 7Li/6Li ratios in the local interstellar
medium (ISM) indicate that the ratio may vary from the
solar system value of 12.3 (Anders & Grevesse 1989), but
the results are far from conclusive. Four different isotope
ratios have been published for the line of sight toward Oph: 25 (Ferlet & Dennefeld 1984) and 6.8þ1:4
1:7 for a single-component fit to the data (Meyer et al. 1993), 1.4þ1:2
0:5
(0.6) and 8:6 0:8 (1.4) (Lemoine et al. 1995), and an
average 7Li/6Li ratio of 13:2 6:3 for their two-component
fit (Howarth et al. 2002). Lemoine et al. (1993) found an isotope ratio of 12.5þ4:3
3:4 toward Oph, consistent with the
solar system value. Knauth et al. (2000) derived isotope
ratios for two velocity components toward o Per of
1:7 0:3 and 3:6 0:6 and an isotope ratio of 10:6 2:9 for
a single component toward Per. The latter ratio is marginally consistent with the value of 5.5þ1:3
1:1 reported by Meyer et
al. (1993). As discussed in x 2, an accurate template for the
velocity components along a given sight line appears to be
required before consistent results emerge.
In our earlier study (Knauth et al. 2000), high-resolution
(R 180; 000) spectra were obtained on the stars o Per and
Per in the Perseus OB2 association. These stars were
chosen for their proximity to IC 348, an active star-forming
region. These data showed that the 7Li/6Li ratio varies from
the solar system value toward Per to a value of approximately 2 toward o Per; o Per resides closer to IC 348 than
does Per and has an order of magnitude higher flux of cosmic rays (Federman, Weber, & Lambert 1996). Knauth et
al. (2000) suggested that the ratio of 2 was clear evidence for
newly synthesized lithium toward o Per resulting from GCR
spallation reactions. These data barely resolve the two
velocity components separated by 3 km s1 toward o Per.
To refine these findings and to probe directions with interstellar clouds separated by 1 km s1, we obtained
269
ultra–high-resolution (UHR) observations toward other
stars in the Perseus OB2 and Scorpius OB2 associations.
The stars in Sco OB2 lie close to an active star-forming
region approximately centered on Oph. Furthermore, Ori was observed because it resides in a region of active star
formation. In addition, an UHR survey of interstellar K i
(Welty & Hobbs 2001) indicated that Li i could possibly be
detected toward Ori. 20 Aql lies relatively far from active
regions of star formation and serves as a useful comparison.
The main goals of the present study were to determine
reliable 7Li/6Li isotope ratios in order to probe the extent of
variations in its value and to help constrain the roots of Li
production. Since 6Li is produced primarily through GCR
spallation reactions, the GCR contribution to 7Li can be
accurately determined. Assuming that the BBN contribution is known from observations of halo stars, the remaining
7Li comes from Population I stars. These assumptions
enabled interstellar constraints to be placed on the stellar
production of 7Li. The organization of this paper is as follows. In x 2 we describe the observations and the data reduction. The profile syntheses for all data sets are discussed and
the results are tabulated in x 3. The isotope ratios, Li i and
K i column densities, elemental abundances, depletion, and
elemental K/Li ratios for each sight line are given in x 4. In
x 5 we discuss implications for future Li studies and thermal
versus turbulent broadening. The results of the 7Li/6Li isotope ratios are discussed in xx 6 and 7. In x 8 constraints on
stellar sources of 7Li are presented. Finally, the results are
summarized and suggestions for future work are contained
in x 9.
2. OBSERVATIONS AND DATA REDUCTION
High signal-to-noise, high-resolution spectra of the
Li i doublet are needed, since coupled to the relatively
weak absorption, the fine-structure separation of the 7Li i
doublet (3p 2 P3=2; 1=2 2s2 S1=2 at 6707.764 Å for J ¼ 32 and
at 6707.915 Å for J ¼ 12) is comparable to the isotope shift
of 6Li (0.160 Å). The result is a blend of the 7Li and 6Li
lines (Ferlet & Dennefeld 1984; Sansonetti et al. 1995). The
situation is further complicated by the existence of several
interstellar clouds along the line of sight (typically six
clouds per kiloparsec). For this reason, UHR surveys
(resolving power R 600; 000) of K i (Hobbs 1974a,
1974b; Welty & Hobbs 2001) were used to find lines of sight
with simple velocity structure (that is, one or two components) with significant amounts [NðK iÞ 1011 cm2] of
absorption. Assuming a constant K/Li ratio (White 1986;
Welty & Hobbs 2001), these surveys yielded lines of sight to
pursue in the quest for interstellar Li i absorption. Table 1
gives the stellar data for the objects studied here.
In order to extract the interstellar 7Li/6Li ratio, knowledge about the velocity structure for a given line of sight
(LOS) is necessary. Data on another species are required for
use as a template of the velocity structure (Lemoine et al.
1993, 1995; Lambert et al. 1998; Knauth et al. 2000;
Howarth et al. 2002). We chose K i because it has a similar
ionization potential and is likely to reside in the portion of
the interstellar cloud containing Li i (White 1986; Welty &
Hobbs 2001). The method used here differs from that of
Lemoine et al. (1993, 1995), who adopted K i 7699 as their
velocity template. Instead, K i 4044 was chosen since it is
of comparable strength to the Li i lines, although 7699 is
still useful in revealing weaker structure that may be visible
270
KNAUTH, FEDERMAN, & LAMBERT
Vol. 586
TABLE 1
Stellar Data
Star
Name
Spectral Type a
HD 5394 .................
HD 23180 ...............
HD 22951 ...............
HD 24534 ...............
HD36861 ................
HD 87901 ...............
HD 116658..............
HD 148184..............
HD 149757..............
HD 179406..............
Cas
o Per
40 Per
X Per
Ori
Leo
Vir
Oph
Oph
20 Aql
B0 IVe
B1 III
B0.5 V
O9.5 V
O8 IIIf
B7 V
B1 III–IV+
B2 Vne
O9.5 V
B3 V
a
b
c
2.1. Ultra–High-Resolution Observations
2.1.1. McDonald Observatory
All observations were taken with the double-pass spectrograph of the 6 foot camera on the 2.7 m Harlan J. Smith
Telescope at the University of Texas McDonald Observatory. Two spectral regions were observed, one centered on
K i 4044 and the other on Li i 6708. The former setting
allowed a wavelength coverage of approximately 1.5 Å,
while the latter yielded 2.5 Å of spectrum. Since such a narrow wavelength range was imaged, care was needed to
ensure that interstellar features with large Doppler motions
with respect to the Earth remained within the range of the
detector. For this purpose, hollow cathode lamps containing Th-Ar and Li, as well as the solar spectrum, were utilized
TABLE 2
Li i and K i Parameters
Species
IPa
(eV)
fsb
(Å)
7Li i ........
5.392
6707.764
0.4946
6707.915
0.2473
6707.924
0.4946
6708.075
0.2474
4044.143a
0.006089
K i..........
a
5.392
4.341
ffsa
hfsb
(Å)
fhfsc
6707.757
6707.769
6707.908
6707.920
6707.922
6707.925
6708.073
6708.076
. . .d
0.18550
0.30910
0.09270
0.15460
0.16490
0.32970
0.08243
0.16490
. . .d
Morton 1991.
Sansonetti et al. 1995. The ‘‘ fs ’’ stands for ‘‘ fine structure ’’ and ’’
hfs ’’ for ‘‘ hyperfine structure.’’
c Welty et al. 1994.
d Hyperfine splitting of K i 4044 was found to be negligible.
b
BV a
(mag)
la
(deg)
ba
(deg)
db
(pc)
E(BV )c
(mag)
2.39
3.86
4.98
6.10
3.39
1.35
1.04
4.42
2.58
5.36
0.10
0.02
0.05
0.29
0.19
0.11
0.13
0.28
0.02
0.09
123.58
160.36
158.92
163.08
195.05
226.43
316.11
357.93
6.28
28.23
2.15
17.74
16.70
17.14
12.00
48.93
50.84
20.68
23.59
8.31
188
453
283
826
324
24
80
150
140
373
0.14
0.26
0.23
0.56
0.12
...
...
0.49
0.32
0.33
Simbad database, operated at CDS, Strasbourg, France.
Perryman et al. 1997.
Papaj et al. 1991.
within the 4044 line profile. Table 2 lists the ionization
potential (IP; Morton 1991), laboratory wavelengths for
Li i and K i (Morton 1991; Sansonetti et al. 1995), and oscillator strengths for Li i and K i fine-structure lines (Morton
1991) used in this study. The oscillator strengths for the Li i
hyperfine transitions were extracted from the fine-structure
oscillator strengths (e.g., Welty, Hobbs, & Kulkarni 1994).
The hyperfine structure of K i 4044 was found to be negligible (0.002 mÅ).
6Li i ........
Va
(mag)
to determine the spectrograph settings. Only a single weak,
uncataloged line appeared in the Th-Ar spectrum at 6708.
Therefore, a Li hollow cathode was used for wavelength calibration, even though the pressure broadened Li lines could
not be used to determine the instrumental width. An echelle
grating was combined with a cross disperser, which was
used in second order for the blue setting and in first order
for the red one. A CuSO4 (RG 610) blocking filter for the
blue (red) setting was placed behind the slit of the spectrograph to isolate the light from a specific order. A slit width
of 145 lm (slit 2) was used. The echelle grating position was
moved each night to minimize the effects of cosmic rays as
well as any flat-field artifacts, in hindsight a very wise decision. Stellar exposures of 30 minutes or less also minimized
the confusion caused by cosmic rays. For many of the
observing runs, a stellar flat from a bright star was obtained.
Unfortunately, because of poor weather conditions, a stellar
flat was not obtained for every observing run. This combination of procedures enabled the detection of 1% CCD
defects and minimized their effect on the data.
The resolution was determined in the course of the analysis by the FWHM of the Th line at 4043.395 Å. The FWHM
was determined to be 0.01172 Å with 7 pixels per resolution
element at this wavelength setting. While a lithium hollow
cathode provided the wavelength calibration for the 6708 Å
region, the lithium lines are dominated by pressure broadening (FWHM 0:0425 Å). Since the same spectrograph was
utilized for both sets of observations and the resolution is
expected to be independent of the wavelength region, the
resolution determined from the K i setting was used for the
Li i setting. At this resolution, the effects of thermal gas
motions must be considered in determining an accurate
instrumental width. After correcting for thermal broadening at an assumed temperature of 300 K, the instrumental
width was found to be 0.833 km s1, which corresponds to
an R of 360,000. This resolution was chosen to discern interstellar clouds separated by 1 km s1.
The reddened stars, o Per, 40 Per, X Per, Oph, and Oph, in the Perseus OB2 and Scorpius OB2 associations
were observed, as were Ori and 20 Aql. Data were also
acquired on the bright unreddened stars Cas, Leo, and Vir. All the observations were obtained over eight observing
runs between 1998 August and 2000 December. Table 3
contains the following information for each observing run:
INTERSTELLAR 7Li/6Li RATIOS
No. 1, 2003
271
TABLE 3
UHR Li i and K i Observations
Star
Species
Observatorya
Date
Exposure Time
(s)
S/N pixel1b
Cas.......................
40 Per......................
Li i
Li i
o Per........................
Li i
McD
McD
McD
McD
McD
McD
McD
McD
McD
McD
McD
McD
McD
McD
McD
McD
McD
AAO
AAO
AAO
McD
AAO
McD
AAO
McD
McD
AAO
McD
AAO
AAO
AAO
McD
McD
McD
McD
McD
1998 Aug 21998 Aug 8
1999 Aug 11–1999 Aug 16
2000 Sep 15–2000 Sep 21
1998 Dec 30–1999 Jan 1
2000 Dec 23–2000 Dec 30
1999 Jul 24–1999 Jul 28
1999 Aug 11–1999 Aug 16
2000 Dec 23–2000 Dec 30
1999 Dec 28–1999 Dec 30
2000 Sep 15–2000 Sep 21
1998 Dec 30–1999 Jan 5
1999 Dec 28–1999 Dec 30
1998 Dec 30–1999 Jan 5
1999 Dec 28–1999 Dec 30
2000 Dec 23–2000 Dec 30
2000 Dec 23–2000 Dec 30
2000 May 11–2000 May 13
1995 Jun 19–1995 Jun 20
1995 Jun 19–1995 Jun 20
1995 Jun 19–1995 Jun 20
2000 Sep 15–2000 Sep 21
1995 Jun 19–1995 Jun 20
2000 May 11–2000 May 13
1994 Jun 10–1994 Jun 12
1998 Aug 2–1998 Aug 8
2000 Sep 15–2000 Sep 21
1995 Jun 19–1995 Jun 20
1999 Jul 24–1999 Jul 28
1995 Jun 19–1995 Jun 20
1995 Jun 19–1995 Jun 20
1995 Jun 19–1995 Jun 20
1998 Aug 2–1998 Aug 8
1999 Aug 11–1999 Aug 16
2000 Sep 15–2000 Sep 21
1999 Jul 24–1999 Jul 28
1999 Aug 11–1999 Aug 16
19,200
14,400
43,200
103,926
81,827
7,200
27,000
28,800
45,000
75,600
52,200
21,600
37,800
10,800
4,500
12,127
13,200
2,700
5,700
27,900
9,000
5,700
50,400
64800
13,500
12,600
4,800
44,700
3,600
2,700
5,800
51,728
43,200
18,000
32,400
55,453
1000
270
470
720
600
100
290
200
150
290
420
370
1050
850
170
470
320
130
100
320
270
160
150
415
500
430
200
580
140
160
130
410
410
250
100
200
Ki
X Per.......................
Li i
Ori .......................
Li i
Leo ......................
Li i
Vir .......................
o Sco .......................
Oph ......................
Oph......................
Ki
Li i
Ki
Ki
Ki
Li i
Ki
Oph ......................
Li i
Ki
HD 154090..............
HD 165024..............
l Sgr .......................
20 Aql .....................
Ki
Ki
Ki
Li i
Ki
a
McD ¼ McDonald Observatory; AAO ¼ Anglo-Australian Observatory.
Final S/N per resolution element from combined binned spectrum: o Per: Li i 3040, K i 1050; 40 Per: Li i 1500;
X Per: Li i 930; Ori: Li i 1440; Oph: Li i 1130, K i 330; Oph: Li i 2600, K i 1500; 20 Aql: Li i 1725, K i 780;
Cas: Li i 2825; Leo: Li i 3675, K i 480; Vir: Li i 1330, K i 900.
b
the stars and species observed, the observatory, the dates of
the observing run, the total exposure time for all observations, and the signal-to-noise ratio (S/N) per pixel. Also
noted in Table 3 are the S/N per resolution element of the
final summed spectra for all stars.
2.1.2. Anglo-Australian Observatory
Additional data on Li i and the strong K i line at 7699 Å
were obtained (Lemoine 1995) from the Anglo-Australian
Observatory (AAO) archive. Li i and K i were observed
toward the stars and Oph, while only K i was observed
toward the stars o Sco, Oph, HD 154090, HD 165024, and
l Sgr. A synopsis also appears in Table 3.
The K i data were reduced to verify the accuracy of our
Li i data reduction. Comparison of the measured values of
W for K i 7699 with those presented in Welty & Hobbs
(2001), based on the same data, show excellent agreement
with the exception of the absorption toward o Sco and HD
154090, which differ by approximately 10%. This overall
agreement lends confidence that our lithium measurements
toward and Oph are reliable.
2.2. Standard Data Reduction
The data were reduced in a standard way utilizing the
NOAO SUN/IRAF software (revision 2.11.3). Dark, bias,
and flat lamp exposures were taken each night to remove
any instrumental effects due to the CCD detector. Comparison spectra were taken periodically throughout the night,
typically every 2 hr. The average bias exposure was subtracted from all raw stellar, comparison (Th-Ar), and flat
images. The scattered light was fitted by a low-order polynomial, in both the dispersion direction and perpendicular to
it for the multiorder observations, and removed. The pixelto-pixel sensitivity was taken into account by dividing the
normalized flat into the stellar spectra. The normalized flat
is the average of 10–20 flat lamp (5–10 stellar flat) exposures,
with each exposure having a flux level 3–5 times that of a
single stellar exposure. Next, the pixels perpendicular to the
dispersion were summed in each order for each stellar and
272
KNAUTH, FEDERMAN, & LAMBERT
comparison lamp exposure. The extracted spectra were
placed on an appropriate wavelength scale using the Th-Ar
or Li comparison spectra and were Doppler-corrected. The
spectra were then co-added and normalized to unity with a
low-order polynomial, yielding a final spectrum with high
S/N (based on the rms of the deviations in the continuum),
approximately 1000 : 1 per resolution element. Over the narrow wavelength range covered by these spectra, there were
at most slight variations in the continua.
Each stellar spectrum was carefully examined for flat-field
artifacts or cosmic rays before the final spectrum was created. Cosmic rays that survived the reduction were
removed. When the cosmic rays coincided with the interstellar features in a spectrum, as was the case with three exposures of Oph, two exposures of 20 Aql, and one exposure
of o Per, the affected spectra were not included in the final
sum. The standard reduction process was successful in
extracting the data for most of the observing runs. While
examining the spectra for cosmic rays, the presence of two
1% features (CCD defects) that were not removed during
the flat-fielding process were detected. An improved flatfielding technique using stellar spectra was applied before
the remaining steps in the extraction process.
The defects were seen in all spectra, but for 60% of the
observing runs the defects were far (200 pixels) from where
the interstellar line(s) appeared. The defects were a problem
for the McDonald observing runs of 1998 August, 1999
December, and 2000 December. The basic solution followed
the prescription for removal of fixed pattern noise in spectra
acquired with the Goddard High-Resolution Spectrograph
(Cardelli & Ebbets 1994). The spectrum of an unreddened
star was used as a template; removal of the defect was
accomplished by dividing the affected spectrum by the template in pixel space, since defects remain fixed in pixel space.
Multiple observing runs of the interstellar species toward
the same star gave an additional check that proved useful in
minimizing the effect of the 1% defects. Where other measures exist for comparison (Meyer et al. 1993; Lemoine et al.
1995; Knauth et al 2000; Howarth et al. 2002), there is good
to excellent correspondence among spectra. The McDonald
spectra were binned by 3 pixels to increase the S/N, a valid
procedure since there are 7 pixels per resolution element,
and the AAO data for and Oph were binned to the lower
resolution of the McDonald data before combining them.
3. PROFILE SYNTHESIS
Each interstellar line profile yields the number of velocity
components, the wavelength at line center, , FWHM, and
W, for each component. The number of velocity components in the Li i spectrum was obtained from a visual examination of the K i spectrum. For symmetric interstellar lines,
the determination of the W is accomplished by fitting a
Gaussian profile to the interstellar feature for comparison
with the results from the profile synthesis. Asymmetric
interstellar lines are telltale signs of multiple components
along a given line of sight. If more than one component was
present, the task of determining W involved simple Gaussian deblending. Unresolved velocity structure is treated as
a single component and may result in slight differences in
the derived quantities. There are additional uncertainties in
the calculated VLSR (0.5 km s1) derived from the K i and
Li i lines. This difference is attributed to the uncertainty in
the wavelength calibration for the Li i region, the result of
Vol. 586
TABLE 4
Li i and K i Upper Limits
Star
Species
Vir ...............
7Li i
Leo ..............
40 Per..............
Ori ...............
Cas...............
Ki
7Li i
Ki
7Li i
7Li i
7Li i
W
(mÅ)
N
(cm2)
0.11
0.18
0.05
0.34
0.11
0.11
0.07
5.60 108
2.04 1011
2.55 108
3.88 1011
5.60 108
5.60 108
3.55 108
there being larger widths for the pressure broadened, hollow
cathode lines of Li i. For the stars, Cas, Leo, and Vir,
as well as 40 Per and Ori, no detection of Li i or K i was
evident. For these stars the 3 upper limits, determined
from the S/N and the FWHM of the instrument function,
are listed in Table 4 and will not be discussed further. Continuum placement uncertainties were not included in the
error budget because variation in the continua was found to
be insignificant.
If a single velocity component is present along a given
sight line, then a simple Gaussian can be fitted to the line
profile in order to determine these three parameters. A more
realistic approximation of the profile function is given by a
Voigt profile. While a Voigt profile more precisely describes
the wings of saturated absorption lines, Gaussian profiles
were adequate in fitting the weak Li i and K i lines observed
in this study. If multiple components are present, the determination of b-value, VLSR, and N is more complicated.
A C++ code (J. Zsargó 2000, private communication)
was utilized to compare the observed K i and Li i line profiles with synthetic ones. The two profiles were synthesized
independently. Comparison of observed and synthetic profiles was performed with a second-order gradient-expansion
Marquardt algorithm (Bevington & Robinson 1992). The
synthetic profiles and 2 were calculated with initial values
of the parameters. The gradient-expansion method adjusts
the initial values and the process is repeated with the
updated values until a change in 2 of less than 0.01%
occurs. An additional feature of the program allowed any
parameter to be fixed in order to arrive at a solution. One or
more parameters were fixed only when a high degree of certainty in the initial parameters was present. For details of
the profile synthesis, see Zsargó & Federman (2002).
Reliable initial values are crucial to finding the most
appropriate minima since there could be several minima
caused by multiple velocity components or lower S/N. The
syntheses gave the b-value, VLSR, and N for each component
as well as the total W. Agreement between the observed W
and that from the combination of the components was at
the 1 level, with the exception of the 7 km s1 component
toward o Per (see x 3.2). The best fit was inferred from Ftests for trials based on fits using different numbers of components. Figures 1–5 show our UHR spectra obtained
toward o Per, X Per, Oph, Oph, and 20 Aql, respectively.
From a visual examination of our spectra, only the clouds
toward o Per reveal large 6Li abundances (compare Figs. 1–
5). Profile synthesis provides a quantitative isotope ratio in
all cases. In addition, the enhanced 6Li abundance toward o
Per is clearly seen in our lower resolution spectrum (Knauth
et al. 2000).
INTERSTELLAR 7Li/6Li RATIOS
No. 1, 2003
273
Fig. 2.—Synthesis of the Li i data toward X Per assumes two velocity
components (S. R. Federman 2001, private communication). See Fig. 1 for
a description of the plot. Residuals (data fit) are offset to 1.01 (dashed
line).
Fig. 1.—Top: Preferred two-component synthesis of the K i data toward
o Per using b-values determined from the line profile. Middle: Twocomponent fit of Li i using b-values determined from 7Li profile. Bottom:
Velocity-structure information from the K i synthesis applied to Li i. In all
plots, the data are represented by the filled circles, the fit is the solid line,
and the dashed lines are the residuals (data fit), offset to 1.02 and 1.08 for
K i and Li i, respectively.
3.1. o Per
At our resolution, a very weak absorption line provided
by a single cloud of pure 7Li would be seen as a resolved
doublet with the ‘‘ blue ’’ line twice the depth of the ‘‘ red ’’
line. Addition of 6Li increases the depth of the ‘‘ red ’’ line,
which reduces the blue-to-red ratio of line depths below the
2 : 1 ratio set by atomic physics (i.e., LS coupling). Superficial inspection of the profiles at the telescope showed that all
lines of sight but one were at or close to this 2 : 1 ratio. The
exception was o Per, where the ‘‘ red ’’ line appeared deeper
than expected (Fig. 1), a telltale signature that either a high
6Li abundance in the cloud or the presence of a second
higher velocity cloud contributing 7Li at about +7 km s1
from the main cloud could be masquerading as 6Li; +7 km
s1 is the isotopic shift. The weaker 6Li line, which is about
0.16 Å to the red of the weaker 7Li line, would be of value in
determining the reason for the deeper ‘‘ red ’’ line, but,
unfortunately, the line is indistinguishable from the noise of
the spectrum (see also the lower resolution spectra published by Knauth et al. [2000]). Figure 6 shows our previous
high-resolution Li i spectrum (Knauth et al. 2000), fit with a
synthetic spectrum assuming a solar system isotope ratio of
12 for both velocity components. Examination of Figure 6
clearly shows an enhanced 6Li abundance.
Our template K i 4044 profiles and those provided by
observations of the K i 7699 line (Welty & Hobbs 2001)
serve to constrain the distribution of clouds toward o Per.
Our 4044 spectrum shows two clouds separated by about 3
km s1, and both clouds are seen to contribute to the Li i
profile (Fig. 1). Welty and Hobbs’s higher resolution observations of the 7699 Å K i line show that each of our clouds is
a pair of clouds with a velocity separation of 1.0 km s1 for
the weaker cloud and 1.3 km s1 for the stronger cloud. A
fifth cloud, at a velocity of about 2 km s1 to the red of the
stronger line in Figure 1, has a column density of only 6% of
that in the stronger cloud. This fifth component is hidden by
the noise in Figure 1.
Independent analyses of the Li i and K i profiles began
with the introduction of the two velocity components
clearly detected in Figure 1. The reduced 2 (2/) of both
fits (Fig. 1) are at the 90%–95% confidence level. Derived
parameters from these independent fits are given in Table 5.
Considering the accuracy of the velocity scales, the small
difference in the cloud velocities derived from the K i and
Li i syntheses is not significant. As can be seen in Figure 6, a
solar system ratio for both velocity components yields a
poor fit to the data. A fit to the Li i absorption was also
made using the VLSR and b-values from the K i synthesis
(see Fig. 1, bottom). Although the isotopic ratios are altered
slightly, the quality of this fit is poor (2 = ¼ 3:73) on
account of the smaller b-values for K i. This suggests (see
below) that thermal broadening is the main contributor to
the b-values.
Our earlier spectra of the K i and Li i absorption were
acquired at a factor of 2 lower resolution and with S/N of
2500 : 1. A two-cloud analysis gave isotopic ratios of
1:7 0:3 and 3:6 0:6 for the 4 and 7 km s1 clouds,
respectively (Knauth et al. 2000). Our result, for the 4 km
274
KNAUTH, FEDERMAN, & LAMBERT
Fig. 3.—Top: One-component synthesis of the K i data toward Oph
using b-values determined from the line profile. Middle: One-component fit
of Li i using b-values determined from 7Li profile. The slight disagreement
for the weaker 7Li line with the results of the synthesis is most likely caused
by a CCD defect within the level of noise. Bottom: Fit of Li i using velocitystructure information from the K i synthesis. See Fig. 1 for a description of
the plots. Residuals (dashed lines) for K i and Li i are offset to 1.03 and 1.01,
respectively.
s1 component, of 2:1 1:1 confirms the earlier result, but
our current result of 8:1 2:1 is in poor agreement with the
earlier result. The measured W values for the blue component of 7Li i and both components of K i agree well with our
previous measurements (Knauth et al. 2000). However, the
red component in Li is 0.1 mÅ larger than that obtained in
Knauth et al. (2000). The larger W could be due to an undetected CCD defect. Therefore, the profile synthesis of the
UHR data was also performed utilizing the results of
Knauth et al. (2000). This synthesis resulted in a reasonable
fit to the data. Although both syntheses are equally acceptable, the resulting column densities only overlap at the 3 level, suggesting that the observational errors alone are
inadequate in representing the true uncertainty. A solar system isotope ratio for both components is unlikely due to the
almost one-to-one correspondence between both members
of the Li i doublet (see Fig. 6).
Vol. 586
Fig. 4.—Top: Preferred two-component synthesis of K i using b-values
from UHRS CN observations (Crawford et al. 1994) toward Oph.
Middle: Synthesis of Li i using b-values determined from 7Li profile.
Bottom: Synthesis of Li i using velocity-structure information determined
from K i. See Fig. 1 for a description of the plots. Residuals (data fit) are
offset to 1.01 (dashed lines) for both K i and Li i.
As noted above, our K i components are each a blend of
two close components. If these four clouds are included
instead of two, constrained fits are possible using the b-values from Welty & Hobbs (2001), although the isotopic
ratios are not changed significantly. Three component syntheses of the K i and Li i line profiles resulted in an increased
2/ (1.51 for K i and 3.32 for Li i) when compared to the
two component syntheses. Application of an F-test (Lupton
1993) shows that a third component is justified at the 70%
confidence level, while a fourth component yields only 50%
confidence. In order to investigate the possibility of 7Li masquerading as 6Li, another synthesis was performed that
incorporated the reddest component from Welty & Hobbs
(2001). This additional component was found to be indistinguishable from the noise in the 4044 Å line profile. Again
the quality of the fit is poor and does not change the occurrence of the low Li isotope ratio.
An additional probe of cloud structure along the line of
sight is provided by the Na D profile (Welty & Hobbs 2001),
No. 1, 2003
INTERSTELLAR 7Li/6Li RATIOS
275
Fig. 6.—We show our previous high-resolution Li i spectrum (Knauth et
al. 2000) with a 7Li/6Li ratio of 12, as a useful comparison. It is quite clear
that a solar system ratio provides a poor fit to the data toward o Per. See
Fig. 1 for a description of the plot. Residuals (data fit) are offset to 1.04
(dashed line).
Fig. 5.—Top: Two-component synthesis of K i toward 20 Aql using bvalues determined from line profile. Middle: Li i synthesis using b-values
determined from 7Li profile. Bottom: Li i synthesis determined from velocity-structure information derived from K i. See Fig. 1 for a description of
the plots. Residuals (dashed lines) for K i and Li i are offset to 1.03 and 1.01,
respectively.
a more sensitive monitor of thin clouds than the 7699 Å line.
Four additional clouds are seen in Na D: two to the red of
the stronger K i line in Figure 1 and two to the blue of the
weaker K i line. Na i column densities in these clouds do not
exceed 1% of the observed column in the strongest of the five
clouds detected in both Na D and K i. Of relevance—
perhaps—to the Li isotopic analysis is that the velocity
shifts of additional clouds to the red are approximately
equal to the velocity shift of 6Li from 7Li in the two main
pairs of clouds.
These clouds of low Na i column density appear at velocities relative to the main clouds that, if they contributed 7Li
absorption, would reduce and even eliminate the need for
6Li absorption from the main clouds. Another synthesis
resulted in a reasonable fit to the data with limits to the
7Li/6Li similar to the solar system value for all velocity components. In order to achieve this, however, the ratio of
N(Li i) to N(Na i) for the two reddest clouds must be about
104 times that in typical clouds (Welty & Hobbs 2001).
Unfortunately, this argument replaces one astrophysical
puzzle—why is the 7Li/6Li ratio so low?—with another—
why is the ratio of Li to Na atoms so high in the higher
velocity clouds? Inclusion of these low column density components into the synthesis adds the additional complication
that these components contain little or no K i absorption. It
is interesting to note that the higher velocity clouds are seen
in both Na i and Ca ii toward o Per, but only Ca ii absorption is detected toward Per for which a normal 7Li/6Li
ratio is found (Knauth et al. 2000). The uncertainties and
substitution of one puzzle for another suggest that the best
results from the profile synthesis of the o Per data are given
by the two-component fit, indicating that the low relative
strengths for the 7Li doublet are the result of enhanced 6Li
in one or both of the clouds.
3.2. Other Sight Lines
The analyses of the other sight lines yielded lower limits
to the 7Li/6Li ratio consistent with the solar system value.
The weaker 7Li line was used to constrain the maximum
amount of 6Li present, since this feature is a blend of the 7Li
and 6Li lines and the ratio of 7Li is 2 : 1. The profile synthesis
produced a formal detection to Oph that is consistent with
these limits. The syntheses for these four sight lines also provided the means to extract the relative importance of
thermal and turbulent broadening from the b-values of Li
and K. We now describe the particulars for each sight line.
X Per.—At least one velocity component is clearly
detected in Li i absorption. Unfortunately, data on K i
4044 are not available for this sight line, a result of X Per
being the faintest star studied here. Observations of K i
7699 reveal more complicated velocity structure (two components with an R 200; 000 [S. R. Federman 2001, private
communication] and four components with an R 600; 000
[D. Welty 2001, private communication]). All three syntheses yield good fits, with 2/ at the 85%–95% confidence
level. An F-test on the data revealed 95% confidence in the
two-component fit and 74% confidence in the four-
276
KNAUTH, FEDERMAN, & LAMBERT
Vol. 586
TABLE 5
Results of Profile Syntheses
Species
VLSRa
(km s1)
b-Value
(km s1)
Wb
(mÅ)
W(FIT)
(mÅ)
Nc
(cm2)
0.09 0.02
0.60 0.02
0.04 0.02
0.08 0.02
0.13 0.03
0.78 0.03
4.7 1.0
30.6 1.1
2.2 1.0
3.8 1.0
1.46 0.35
8.96 0.35
0.63 0.06
0.28 0.06
0.10
0.13
32.0 3.1
14.2 3.1
5.1
6.8
0.86 0.07
0.11 0.07
1.08 0.10
44.0 4.1
5.8 3.6
12.7 1.2
0.41 0.02
0.22 0.02
0.03
0.02
0.47 0.03
0.25 0.03
21.0 1.0
11.4 1.0
1.8
1.0
5.36 0.35
2.89 0.34
0.67 0.04
0.25 0.04
0.07
0.04
1.34 0.04
0.56 0.04
34.0 2.1
12.8 2.1
3.6
2.3
15.6 0.5
6.40 0.47
7Li/6Li
2/
2.1 1.1
8.1 2.1
1.28
1.28
1.28
1.28
1.80
1.80
o Per
7Li i ...................
6Li i ...................
K i.....................
4.19
7.38
4.19
7.38
3.96
6.98
0.53
2.53
0.53
2.53
0.30
0.92
0.12 0.02
0.47 0.02
0.14 0.03
0.74 0.03
X Per
7Li i ...................
6Li i ...................
5.45
7.85
5.45
7.85
1.20
1.15
1.20
1.15
0.63 0.06
0.35 0.06
6.3
2.1
1.12
1.12
1.12
1.12
Oph
7Li i ...................
6Li i ...................
K i.....................
0.887
0.887
0.30
1.27
1.27
0.46
0.93 0.07
1.03 0.10
7.6 4.8
1.30
1.30
1.79
11.7
12.0
1.81
1.81
1.81
1.81
3.62
3.62
9.4
5.6
1.12
1.12
1.12
1.12
1.66
1.66
Oph
7Li i ...................
6Li i ...................
K i.....................
0.620
0.420
0.620
0.420
1.25
0.21
0.65
0.57
0.65
0.57
0.55
0.45
0.41 0.02
0.14 0.02
0.42 0.03
0.31 0.03
20 Aql
7Li i ...................
6Li i ...................
K i.....................
2.06
3.35
2.06
3.35
1.60
3.21
1.06
1.46
1.06
1.46
0.75
0.79
0.67 0.04
0.24 0.04
1.30 0.04
0.59 0.04
a Differences in velocities (0.5 km s1) between Li i and K i lines, toward the same star, are attributed to
uncertainties in using the Li hollow cathode to determine wavelength calibration.
b Measured W with 1 observational uncertainty; No measured W for 6Li are reported.
c Note that N(Li i) is in units of 108 cm2 and N(K i) is in units of 1011 cm2.
component fit. Higher S/N data are necessary to distinguish
the most appropriate velocity structure in Li i absorption for
this line of sight. The best fit, with two velocity components,
is displayed in Figure 2; the results appear in Table 5. The bvalues, 1.20 and 1.15 km s1, are similar to those derived
from K i 7699 absorption, 1.7 and 1.2 km s1 (S. R.
Federman 2001, private communication), suggesting that
turbulence dominates the interstellar clouds toward X Per.
Oph.—Only one velocity component is present. Multiple velocity components are seen in UHR K i 7699 spectra
(Welty & Hobbs 2001). An unrealistic fit resulted (e.g., an
emission line) from a two-component synthesis of the K i
4044 absorption. A significantly weaker third velocity
component is at least 2 km s1 away (Welty & Hobbs 2001)
but within the noise level in our data. The single-component
synthesis of the line profiles is shown in Figure 3. The slight
disagreement for the weaker 7Li line with the results of the
synthesis is most likely caused by a CCD defect within the
level of noise. The 2/ for the top two fits are at the 90%–
95% confidence level. The results of these independent
syntheses are presented in Table 5. The synthesis of Li i with
the smaller b-values derived from the K i fit is not acceptable
(2 = ¼ 2:06); thermal broadening dominates this interstellar cloud. As a test for a single primary component, a
self-consistent b-value was calculated for K i 7699 utilizing
N(K i) derived from 4044 and the total W of the two
velocity components of 7699 (Welty & Hobbs 2001). This
results in a b ¼ 0:75 km s1, which agrees with the b-value
of Welty & Hobbs (2001) within their quoted 0.1 km s1
uncertainty.
Oph.—Two velocity components, separated by 1 km
s1, are clearly detected in both K I and Li i absorption
toward Oph. Figure 4 shows the results of this twovelocity component fit to K i (top) and Li i (middle). There is
also a slight mismatch between the fit and the data in the
weaker 7Li line, which also occurred for Oph, most likely
caused by a CCD defect within the level of noise. The results
of the two component fit are exceptional for K i and quite
good for the Li i absorption, as shown in Table 5. A single
component with an equivalent width of 0.638 mÅ was
INTERSTELLAR 7Li/6Li RATIOS
No. 1, 2003
reported toward Oph by Meyer et al. (1993), 0.55 mÅ for
the combined two components by Lemoine et al. (1995),
and 0.68 mÅ for the combined two components of Howarth
et al. (2002). These measurements agree well with the combined W for the two velocity components, 0.55 mÅ,
reported here. The K i parameters fit the Li i profile as well,
suggesting that turbulent broadening dominates. UHR data
(Welty & Hobbs 2001) show three additional relatively
weak K i 7699 absorption components [NðK iÞ 1 1010
cm2], approximately 1 km s1 to longer wavelengths and 1
and 4 km s1 to shorter wavelengths than the two stronger
components. These relatively weak components are at the
level of the noise in our 4044 spectrum. Inclusion of a third
component into our K i and Li i profile syntheses, with the
parameters taken from Welty & Hobbs (2001), results in
only a 50% confidence in there being another component.
Howarth et al. (2002) also find no evidence for a third
component.
20 Aql.—There are two components clearly detected in
K i and 7Li i. Figure 5 shows the fits to both K i and Li i
using each species to determine the profile parameters. The
results of the independent fits are shown in Table 5. The bottom graph shows the fit to Li i using the parameters derived
from the K i synthesis. This fit has a slightly larger reduced
2 (1.36) because the derived b-values from the Li i fit are
slightly larger (0.5 km s1) than the values found from the
K i fit, indicating the importance of thermal broadening.
4. ANALYSIS
4.1. 7Li/6Li Ratios
The 7Li/6Li ratios, a direct result of the profile synthesis,
are listed in Table 5. The uncertainties were determined
from the weighted average of both the 7Li and 6Li column
densities (when present). The lower limits are derived from
the expected relative line strengths for the 7Li doublet (2 : 1).
Toward o Per, one velocity component is nearly identical to
the value predicted by models of GCR spallation reactions
(2), while the other component may be similar to the solar
system value of 12.3 (see Knauth et al. 2000) as are the isotope ratios, or their limits, toward X Per, Oph, Oph, and
20 Aql.
Other measures of the isotope ratio are available for
comparison. The determination of 7Li/6Li ¼ 5:5þ1:3
1:1 toward
Per, from spectra with R 200; 000 and S/N of 2100
(Meyer et al. 1993), is marginally consistent with our previous observations taken at a similar resolution and with a
S/N of 2500, which resulted in a 7Li/6Li ratio of 10:6 2:9
(Knauth et al. 2000). Toward Oph, where a single velocity
component is present, a ratio of 12.5þ4:3
3:4 was reported
(Lemoine et al. 1993) from spectra obtained at R 100; 000
and S/N of 2700. A single component toward Oph yields
7Li/6Li ¼ 6:8þ1:4 from spectra with R 200; 000 and S/N
1:7
of 2300 (Meyer et al. 1993), but it is actually two velocity
components separated by 1 km s1 (Lambert, Sheffer, &
Crane 1990; Crawford et al. 1994; Crawford & Barlow
1996; Welty & Hobbs 2001; this work). Lemoine et al.
(1995) reported isotope ratios of 1.4þ0:8
0:5 (0.6) and 8:6 0:8
(1.4) for two velocity components separated by 5 km s1
from spectra taken at R of 100,000 and with a S/N of 7500.
The latter is the single component described by Meyer et al.
(1993). In a recent paper, Howarth et al. (2002) report an
average 7Li/6Li ratio of 13:2 6:3 for their two-component
277
fit toward Oph based on spectra taken at R 106 and a
signal-to-noise ratio of 1200 : 1.
Lemoine et al.’s results for the weaker bluer component
at Vhelio ¼ 19 km s1 with the low isotope ratio toward
Oph raises a number of issues. First, the 7Li i feature is but
a 2–3 detection, and similar positive excursions appear in
their spectrum (their Fig. 6). Second, this component has a
K/Li ratio much smaller than that found here for this sight
line and others. Another problem is that there is no molecular material associated with the velocity component having
a low isotope ratio. For all other lines of sight in which Li i
has been detected, substantial molecular material is found
at the velocity of Li i 6707. The component may be the
result of using the strong K i line at 7699 Å as the velocity
template because it detects weaker interstellar features than
will be seen in Li i. The weak K i 7699 component is at the
limit of our S/N in 4044. No evidence for the weak velocity
component in Li i toward Oph is suggested by our data or
that of Howarth et al. (2002). These facts seem to indicate
that a velocity component with a low isotope ratio is not
present toward Oph. It is important to note that the S/N
of Lemoine et al.’s data is a factor of 2 larger than the S/N
of our data.
4.2. Neutral Columns of Li and K
Figure 7 shows a comparison of N(Li i) to N(K i). The
least-squares fit (solid line) for all lines of sight studied here
and in Welty & Hobbs (2001) is shown. The dotted line is
the best fit from Welty & Hobbs (2001). Our least-squares
fit includes the uncertainties from this work and Knauth et
al. (2000), as well as the observational uncertainties for Li i
and an assumed 20% uncertainty in the K i values (Welty &
Hobbs 2001 and references therein). Since a fit of data
points derived from individual velocity structure (asterisks)
has a higher correlation coefficient (r ¼ 0:92) than does the
fit (r ¼ 0:81) based on total columns ( filled circles), use of
individual velocity components along the line of sight is
the more appropriate means for comparison. Another
Fig. 7.—Plotted are the values of log½NðK iÞ vs. log½NðLi iÞ for the
lines of sight studied here (asterisks) and from Welty & Hobbs 2001 ( filled
circles). The dotted line is the best fit from Welty & Hobbs (2001), and the
solid line is the least-squares fit to all the data. A typical error bar is shown
in the upper left of the graph.
278
KNAUTH, FEDERMAN, & LAMBERT
consideration is that some of the Li i observations from the
literature are only modest detections at best; higher quality
data may also improve the correlation coefficient. Significantly greater dispersion about our fit arises when alternative component structures discussed in x 3 are included; we
consider this further evidence for the preferred results given
in Table 5.
The furthest data point from the fits is for the LOS
toward HD 154368. This LOS has the largest column of Li i
reported (Snow et al. 1996), almost an order of magnitude
larger than toward Oph, the next largest reported column
density (White 1986; Lemoine et al. 1993; D. Knauth et al.,
unpublished). The measurement toward HD 154368 could
be as much as a factor of 5 too high (J. Black 2002, private
communication), or K i may not scale linearly with Li i.
4.3. Elemental Abundances
The derivation of the total interstellar abundance requires
knowledge of the abundance of Li ii and the amount of
depletion onto grains. In interstellar space, atoms of Li are
predominantly singly ionized in view of the fact that the
ionization potential of Li i is relatively low (5.31 eV). The
ionization potential of Li ii is 75.6 eV. We consider the gasphase abundance here and wait to discuss implications for
depletion onto grains (x 5.2). An adequate estimate of the Li
abundance is obtained through ionization balance and an
independent determination of the electron density (ne). The
electron density is inferred directly from the column of C+,
the most abundant element providing electrons, the gas density (n) (Federman et al. 1994; Knauth et al. 2001), and the
total proton column density [Ntot ðHÞ ¼ NðH iÞ þ 2NðH2 Þ]
(Savage et al. 1977; Bohlin, Savage, & Drake 1978; Diplas &
Savage 1994) along the LOS. Table 6 contains the information on N(H i), N(H2), Ntot(H), and n. Since no precise
interstellar measurements of C+ exist for the clouds toward
o Per, Oph, and 20 Aql, the weighted mean interstellar
ratio of NðCþ Þ=Ntot ðHÞ ¼ ð1:42 0:13Þ 104 (Sofia et al.
1997) was utilized. Precise measures of N(C+)/Ntot(H) are
used in the analysis toward X Per [ð1:06 0:38Þ 104 ;
Sofia, Fitzpatrick, & Meyer 1998] and toward Oph
[ð1:32 0:32Þ 104 ; Cardelli et al. 1993]. Welty & Hobbs
(2001) discussed the effects of large molecules on ionization
balance and found that depletion (or elemental abundances)
of Li, Na, and K are not altered appreciably.
TABLE 6
Hydrogen and Space Densities
Star
N(H i)a
(1020 cm2)
N(H2)b
(1020 cm2)
Ntot(H)c
(1021 cm2)
nd
(cm3)
40 Per........
o Per..........
X Per.........
Ori .........
Oph........
Oph ........
20 Aqlf ......
11.0 4.7
6.61 1.38
5.37 0.75
6.03 2.87
17.0 3.6
4.90 1.14
17.0 2.6
2.88 1.23
4.07 1.44
11.0 3.0e
0.13 0.6
4.27 1.82
4.47 1.02
6.55 0.98
1.68 1.01
1.48 0.61
2.74 0.84
0.63 0.42
2.55 1.21
1.38 0.45
3.01 0.64
...
800
1000
...
400
400
850g
a
Bohlin et al. 1978; Diplas & Savage 1994.
Savage et al. 1977.
c N ðHÞ ¼ NðH iÞ þ 2NðH Þ.
tot
2
d Federman et al. 1994.
e Mason et al. 1976.
f Hanson et al. 1992; assumes 15% error.
g Knauth et al. 2001.
b
Vol. 586
Through ionization balance we can determine the gasphase lithium abundance (Li/H), without a correction for
depletion onto grains, via
NðLi iÞ
GðLi iÞ
þ 12 :
ð1Þ
Ag ðLiÞ ¼ log
Ntot ðHÞ ðLi iiÞne
In this equation, Ag (Li) is the elemental abundance, (Li ii)
is the radiative recombination rate constant (Péquignot &
Aldrovandi 1986), G(Li i) is the photoionization rate from
the ground state corrected for attenuation by dust grains.
The theoretical determination of (Li ii) depends on the
photoionization cross section from all levels and is known
to better than 2.5% (Péquignot & Aldrovandi 1986). The
photoionization rates at the cloud surface were determined
from the measured ionization cross sections of Hudson &
Carter (1965a, 1967a) and the average interstellar radiation
field of Draine (1978). A similar expression holds for Ag(K),
where K i rates were determined from the measured ionization cross sections of Hudson & Carter (1965b, 1967b),
Marr & Creek (1968), and Sandner et al. (1981).
The photoionization rates need to be corrected for the
effect of attenuation from dust in the interstellar cloud(s).
This is accomplished through use of extinction curves for
the particular line of sight. If one were to assume that all
dust particles are the same throughout the Galaxy, one
would get an average Galactic extinction curve (Code et al.
1976).
From the extinction curves (Papaj, Krewlowski, &
Wegner 1991), the value of selective extinction, A, for the
reddened star, can be derived with the equation
A ¼
Eð V Þ
EðB V Þ þ AV :
EðB V Þ
ð2Þ
Equation (2) assumes that there is no interstellar extinction along the line of sight toward the unreddened star. This
is generally not the case; therefore we need to include terms
for selective extinction for the unreddened star, A0 and
AV0 . These values of selective extinction were obtained from
the average Galactic extinction curve (Code et al. 1976)
scaled to the E(B V ) of the unreddened star, typically 0.02
mag.
Correcting the photoionization rates for attenuation
involves a multiplicative factor of the form expð- Þ,
where ¼ A /2. The factor of 2 arises because radiation
impinges from both sides of the slab representing a
cloud. An optical depth ( 1200) was chosen for Li i, since
1200 Å lies in the middle of the range in wavelength that
leads to its photoionization. Two optical depths were
used for K i: 1200 for the overlap region and 2500 for
wavelengths longer than 2400 Å, which do not contribute
to the photoionization of Li i. The total attenuated photoionization rate for potassium comprises the sum for the
two intervals. The long-wavelength interval contributes
about 5% to the total. For most of the observed lines of
sight, the extinction curves are similar to the standard
extinction curve (Code et al. 1976), the exceptions being
X Per and Oph. These anomalous extinction curves
arise because X Per is a member of an X-ray binary system and Oph is an emission line star. The different
extinction laws change the photorates by approximately
20%–30% compared to the standard extinction curve;
larger excursions result for the anomalous extinction
curves of X Per (70%) and Oph (40%).
INTERSTELLAR 7Li/6Li RATIOS
No. 1, 2003
TABLE 7
Li and K Abundances and Depletion
Star
Li/H
D(Li)
(dex)
K/H
D(K)
(dex)
o Per.........
X Per........
Oph.......
Oph .......
20 Aql ......
(2.8 1.3) 1010
(6.3 2.5) 1011
(2.4 1.3) 1010
(4.0 1.5) 1010
(1.5 0.3) 1010
0.9
1.5
0.9
0.7
1.2
(2.0 0.6) 108
(3.3 1.0) 109
(1.7 0.8) 108
(2.6 0.9) 108
(1.7 0.4) 108
0.8
1.6
0.9
0.7
0.9
The derived interstellar lithium abundances, shown in
Table 7, are quite similar to other recent determinations
with the exception of X Per and 20 Aql. The LOS toward
X Per exhibits almost an order of magnitude lower Li abundance than other sight lines, while the Li abundance toward
20 Aql is about a factor of 2 less. Using the above analysis
for consistency, previous measures are 2:9 1010 toward o
Per (Knauth et al. 2000), ð3:3 3:6Þ 1010 toward Per
(Meyer et al. 1993; Knauth et al. 2000), 2:7 1010 toward
Oph (Lemoine et al. 1993), and ð3:9 4:9Þ 1010 toward
Oph (Meyer et al. 1993; Lemoine et al. 1995). These values
are different from those presented in the referenced work
because they include our correction for the attenuation of
the radiation field and use a different photoionization rate.
4.4. Depletion onto Grains
The determination of the elemental Li abundance
requires an uncertain correction for the amount of depletion
onto grains. From the gas-phase Li/H abundance, the
amount of Li depletion onto grains was estimated. The
amount of depletion is measured with respect to the solar
system value from meteorites, log½ðLi=HÞ =ðLi=HÞ
,
where ðLi=HÞ
¼ 20:5 1010 (Anders & Grevesse 1989).
The K depletion was derived in a similar manner using
the value of ðK=HÞ
¼ 1:26 107 (Anders & Grevesse
1989). One problem with the use of the solar system values
for depletion measures is the assumption that the solar system value represents undepleted gas in the solar neighborhood. There is evidence (Wielen, Fuchs, & Dettbarn 1996;
Wielen & Wilson 1997) that the Sun actually formed closer
to the center of the Galaxy, by about 2 kpc. If our Sun
formed in a more metal-rich environment, then the comparison with the solar system values is inappropriate. Furthermore, the solar system values represent the interstellar gas
at the epoch of formation, 4.6 Gyr ago.
The amounts of depletion, given by the depletion indices
D(Li) and D(K), are listed in Table 7. The indices for Li and
K are similar for most sight lines (0.9 dex), with the
exceptions of those toward X Per and 20 Aql, which are discussed in more detail below. Welty & Hobbs (2001) also
reported similar amounts of depletion for the two species,
but for a larger number of sight lines. Their depletion indices are DðKÞ DðLiÞ 0:6 dex. The use of a constant
index for Li and K just scales the gas-phase abundance without revealing any new information; direct comparison to
the solar system may be more appropriate in studying effects
of depletion onto grains.
4.5. Elemental K/Li Ratio
Through the comparison of N(Li i) and N(K i) (see our
x 4.2), Welty & Hobbs (2001) suggest a constant elemental
279
K/Li abundance ratio. It is straightforward to calculate the
elemental K/Li abundance ratio and compare it with the
solar system value on a component-by-component basis for
our results. Since the amount of depletion onto grains is
similar for the two species (Welty & Hobbs 2001) for most
directions, its uncertainty in equation (1) can be minimized
by taking the ratio of N(K i) to N(Li i). This ratio, which
also eliminates ne and Ntot(H), is represented by
Ag ðKÞ
NðK iÞ ðLi iiÞ GðK iÞ
;
ð3Þ
¼
Ag ðLiÞ
NðLi iÞ ðK iiÞ GðLi iÞ
where the observed values were compared to the solar system value for Ag (K)/Ag(Li) of 61.6 (Anders & Grevesse
1989). Using the theoretical ratio for ðK iiÞ=ðLi iiÞ
¼ 0:58 (Péquignot & Aldrovandi 1986), the K/Li ratio
then mainly depends on the photoionization rates for K i
and Li i.
For the best-fit velocity structure (see Table 5), the calculated elemental K/Li abundance ratios are consistent with
the solar system value of 61.6 (Anders & Grevesse 1989),
considering their mutual uncertainties (see Table 8). On the
other hand, both significantly higher and lower abundance
ratios result when other component structures are considered. We, therefore, suggest that the K/Li ratio can be used
to discriminate between syntheses based on different velocity structure (see also Steigman 1996). For the LOS toward
X Per, both Li and K abundances are low by a comparable
amount and thus the K/Li ratio is unaffected. This suggests
that the amount of depletion onto grains is substantially
greater or that ne has been underestimated. The one exception is the LOS toward 20 Aql; both components yield elemental K/Li abundance ratios that are about 2 times larger.
As an additional check for this line of sight, our previous
high-resolution data on K i 7699 (Knauth et al. 2001) were
reanalyzed with the velocity component structure deduced
from our UHR data. This reanalysis yielded N(K i) and elemental K/Li abundance ratios similar to those obtained
from 4044. Therefore, the disparate K/Li ratios toward 20
Aql appear to be real. The discrepancy could be the result of
an increase (decrease) in the photoionization rate of Li (K),
of evidence for enhanced depletion of Li compared to K, or
of the possibility that Li production (destruction) was lower
(higher) along this LOS. The low K/Li ratios toward 20 Aql
are discussed further in x 6.
5. INTERSTELLAR MATTERS
5.1. Detecting Lithium
At our resolution, velocity structure separated by 1 km
s1 can be resolved (e.g. Oph). Most of the additional
velocity structure detected in K i 7699 (Welty & Hobbs
2001) is separated by more than 1 km s1 and should be resolved in our data. This situation is reminiscent of the component structure utilized by Lemoine et al. (1995) for the
gas toward Oph. Because its f-value is much larger, K i
7699 reveals components too weak to see in absorption
from either K i 4044 or Li i 6707. Although 7699 is not
an ideal template for velocity structure, it plays an important role in the search for Li i by revealing sight lines where
Li i may be detectable.
In addition to absorption from K i, another useful means
for determining lines of sight having detectable amounts of
Li i is the amount of molecular material present, i.e., CH,
280
KNAUTH, FEDERMAN, & LAMBERT
Vol. 586
TABLE 8
K/Li Abundance Ratios
Star
Number of Components
o Per.................
2
3.96
6.98
3.96
6.70
8.00
4.10
6.83
8.12
9.10
5.85
5.45
7.85
4.69
5.60
7.47
8.49
0.30
0.62
0.42
0.62
0.42
1.65
1.60
3.20
3
4
X Per................
1
2
4
Oph...............
Oph ...............
1
2
3
20 Aql ..............
VLSR
(km s1)
2
N(K)a
(1011 cm2)
1.46 0.35
8.96 0.35
1.47 0.35
4.41 0.35
2.09 0.35
1.47 0.35
4.41 0.35
2.09 0.35
0.39 0.08b,c
30.7 2.7
7.37 0.22d
3.53 0.12d
1.70 0.34c,d
4.60 0.92c,d
4.90 0.98c,d
0.18 0.04c,d
12.7 1.2
5.36 0.35
2.89 0.34
4.09 0.82b,c
2.72 0.54b,c
0.11 0.02b,c
15.6 0.5
6.40 0.47
N(7Li + 6Li)a
(108 cm2)
A(K)/A(Li)
6.90 2.0
34.4 2.1
10.8 2.0
12.1 2.0
11.2 2.3
10.7 2.0
12.7 2.0
7.6 2.0
5.9 2.0
44.2 3.1
37.1 3.1
21.0 3.1
10.5 3.1
20.1 3.1
12.9 3.1
7.3 3.1
49.8 7.7
22.8 1.0
12.4 1.0
19.2 1.1
10.5 1.0
0.1
37.6 2.1
15.1 2.1
58.8 22.1
72.4 5.2
38.0 11.5
101.7 18.6
52.1 13.8
38.3 11.6
97.2 17.2
76.3 23.7
18.5 7.3
195.0 21.9
55.8 5.0
47.2 7.2
45.5 16.2
64.2 16.2
106.6 33.3
69.2 32.5
70.7 12.8
66.0 5.2
65.7 9.4
59.8 12.5
72.8 16.0
237.5
115.8 7.5
118.3 18.6
a
The error is 1 observational uncertainty.
Welty & Hobbs 2001.
c Assumes a 20% uncertainty in measurement.
d Two component: S. R. Federman 2001, private communication; four component: D. Welty 2001, private
communication.
b
C2, or CN. Atoms with low IPs and molecules exist only in
the relatively denser regions of diffuse gas. The molecular
column densities for the lines of sight studied here are presented in Table 9. For sight lines where Li i is found, significant amounts of molecular absorption are also observed
[e.g., NðCHÞ NðC2 Þ 1013 cm2 ; NðCNÞ 1012 cm2 ].
Since 40 Per and Ori are in regions of active star formation
and have relatively strong K i 7699 absorption, they were
considered potential targets. However, no Li i was detected;
of all stars observed in our survey, 40 Per has the lowest
values for N(C2) and N(CN) and Ori has the lowest value
of N(H2) (see Table 6). Therefore, the amount of molecular
material present along a given LOS seems a superior indicator for detectable amounts of Li i.
TABLE 9
Molecular Column Densities
Star
N(CH)
(cm2)
N(C2)
(cm2)
N(CN)
(cm2)
o Pera ................
40 Perb ..............
X Pera ...............
Ori .................
Opha ..............
Opha ...............
20 Aqld ..............
1.3 1013
1.2 1013
3.1 1013
...
3.4 1013
2.5 1013
2.0 1013
2.7 1013
3.6 1012
5.3 1013
...
3.5 1013
1.79 1013c
5.2 1013
2.6 1012
6.4 1011
8.4 1012
...
1.3 1012
2.6 1012
4.2 1012
a
Federman et al. 1994.
B.-G. Andersson 2000, private communication.
Lambert et al. 1995.
d Knauth et al. 2001.
b
c
5.2. Thermal versus Turbulent Broadening
Atoms of K i and Li i seem to reside in the same portion
of interstellar clouds. Through measurements of the b-value
for lines of each species, the kinetic temperature, Tk, and the
turbulent velocity, vturb , can be extracted,
b2 ¼
2kTk
þ 2v2turb :
m
ð4Þ
The parameters k and m are the Boltzmann constant and
the mass of the atom observed. The measured b-values for
each species for individual velocity components along a
given LOS are listed in Table 5. The b-values for Li are generally larger than those for K, with the exception of the LOS
toward X Per and Oph. If vturb is assumed to be negligible,
an upper limit can be placed on the kinetic temperature. The
upper limit on Tk for all sight lines studied here ranges from
100 to 2700 K, values not unexpected for diffuse interstellar
clouds. The kinetic temperature and turbulent velocity can
be determined simultaneously, since there are two independent measures of the b-value. The calculated values of Tk and
vturb for each velocity component are given in Table 10. The
uncertainties are based on an assumed 30% error in the
b-values for both species.
In order to arrive at a solution for vturb for the second component toward o Per, both components toward X Per, and
the single component toward Oph, extreme values of the bvalue (1 ) were utilized. No error bars are quoted for Tk
or vturb for these components. The derived temperatures for
the two components toward o Per are 98 42 and 900 K and
those toward X Per are 520 and 788 K. For the single
INTERSTELLAR 7Li/6Li RATIOS
No. 1, 2003
TABLE 10
Kinetic Temperature and Turbulent Velocity
Star
o Per..................
X Perd ...............
Ophd ..............
Oph ................
20 Aql ...............
Tka
(K)
vturb b
(km s1)
vturb c
(km s1)
98 42
900vturb d
520
788
220
61 27
63 27
287 124
770 330
0.16 0.10
0.70d
0.77
0.43
0.37
0.37 0.23
0.30 0.18
0.47 0.29
0.39 0.24
0.15 0.06
1.20d
1.05
1.00
0.53
0.30 0.13
0.21 0.09
0.67 0.29
0.97 0.42
a
Tk derived from both b-values.
Uses calculated Tk.
c Assumes T = 100 K.
k
d Extreme values of the b-values ( 1 ) were used to arrive
at a solution for Tk and vturb ; no error bars are given.
b
component toward Oph, a temperature of 220 K is
inferred. The two components toward Oph yield temperatures of 61 27 and 63 27 K, and toward 20 Aql the temperatures are 287 124 and 770 330 K. The current
measurements agree with previous determinations of Tk, to
within the assumed uncertainties. From H2 excitation, Savage et al. (1977) deduced temperatures of 48 K toward o Per,
46 K toward Oph, and 54 K toward Oph. From C2 excitation and a simple chemical model for CH, C2, and CN
(Federman et al. 1994), temperatures of 40 K toward o Per,
20 K toward X Per, 60 K toward and Oph (Federman et
al. 1994), and 50–60 K toward 20 Aql (Federman, Strom, &
Good 1991; Hanson, Snow, & Black 1992; Knauth et al.
2001) are derived. From an analysis of C2 excitation,
Wannier et al. (1999) derived a temperature of 40 20 for
the LOS toward o Per and Lambert, Sheffer, & Federman
(1995) derived a temperature of 20–80 K for the interstellar
clouds toward Oph. Howarth et al. (2002) found Tk of
500 130 K and 830 125 K toward Oph. Their higher
temperatures are primarily due to their larger derived
b-values for Li i.
The turbulent velocities were determined for each velocity
component as well as Tk. Our calculated values of vturb for
both interstellar clouds toward Oph are 0:38 0:20 and
0:30 0:16 km s1. These values agree well with previous
þ0:37
1
determinations, 0.39þ0:04
0:10 and 0.330:16 km s , based on
UHR spectra of CH and CN (Crawford et al. 1994). This
agreement with an independent measure places additional
confidence in vturb obtained for the other sight lines. With
the exception of the second component toward o Per and
the LOS toward X Per, the turbulent velocities are substantially less than 1 km s1. For further comparison, vturb was
calculated assuming Tk ¼ 100 K, a typical value for diffuse
clouds. A comparison between vturb and the sound speed
shows that the turbulence is subsonic for all sight lines. The
inclusion of molecular gas decreases the sound speed
slightly and in some instances may lead to sonic turbulence.
6. INTERSTELLAR 7Li/6Li RATIOS: THE RULE
With the exception of one cloud along the line of sight to
o Per and possibly the line of sight toward Oph, our lower
limits and the few other published 7Li/6Li ratios are consistent with an expectation that interstellar Li should be simi-
281
lar to the solar system ratio of 7Li/6Li ¼ 12:3 (Anders &
Grevesse 1989). This expectation is based on observational
evidence that young stars and interstellar gas have a composition quite similar to that of the Sun for elements whose
synthesis is traceable to the same sites likely to control the
7Li and 6Li input into the interstellar medium. Beryllium
and 6Li are both products of spallation between Galactic
cosmic rays and interstellar nuclei. The stellar abundances
of Be in local stars younger than the Sun are very similar to
the solar system abundance (Boesgaard et al. 2001). Oxygen, a product of SN IIs that may also make 7Li by the process, has very similar abundances in the Sun, young
stars, and the interstellar medium (Allende Prieto, Lambert,
& Asplund 2001). Carbon, partly a product of AGB stars
that may also synthesize 7Li, also shares an abundance similar to that of the Sun, young stars, and the local interstellar
medium (Allende Prieto et al.2002). Another line of evidence to support the expectation of similar isotopic ratios
among the clouds and the solar system is that stars of solar
metallicity show very similar abundance ratios for elements
created by different nucleosynthetic processes despite differences in age and birthplace (Edvardsson et al. 1993; Reddy
et al. 2003).
The expectation applies to the total Li abundance too.
Since about 90% of interstellar Li is depleted onto or into
grains, the K/Li ratio is considered. Identification of this
ratio with the elemental ratio assumes depletion of Li and
K, both alkali atoms, is similar. Table 8 shows that seven of
the nine clouds have a K/Li ratio close to the solar system
ratio of 61.6: the mean of the seven clouds is 62.4, and individual measurements are each consistent with the solar system value to within their mutual uncertainties. Our previous
observations toward Per (Knauth et al. 2000) yield a K/Li
ratio of 52:6 3:4. The average of eight additional K/Li
ratios from Welty & Hobbs (2001) is 59, assuming that a
standard extinction curve applies for all sight lines. With the
exception of 20 Aql, all K/Li ratios are consistent with the
solar system value.
The interstellar clouds, toward stars studied here, have
depletion indices ranging from 0.7 to 1.5 dex (assuming
solar Li and K abundances). The sight line with the largest
depletion, X Per, is the most reddened direction. The other
two clouds, the pair toward 20 Aql, have K/Li ’ 116, a
value about twice the solar system ratio. Several speculations may be offered for the anomalous ratios toward 20
Aql: (1) the spectrum of the ionizing radiation is different
from that assumed, (2) Li (relative to K) is more depleted
onto grains, or (3) the gas may have been mixed with
ejecta—now cold—from the supernova thought to be
responsible for Radio Loop I (Hayakawa et al. 1979;
Sembach, Savage, & Tripp 1997). Factors of 2 spreads in
abundance ratios are not uncommon; Welty & Hobbs
(2001) discuss a collection of ‘‘ discrepant ’’ clouds toward
the Sco-Oph region.
7. INTERSTELLAR 7Li/6Li RATIOS: THE EXCEPTION
One of the clouds toward o Per has an exceptional 7Li/6Li
ratio of 2:1 1:1. This is clearly shown by spectra taken
with two different spectrographs. Our previous lower
resolution data shown in Figure 6 clearly show that a solar
system ratio cannot apply to either interstellar cloud toward
o Per. From a visual examination of the data, both components are in almost a 1 : 1 correspondence and not in the 2 : 1
282
KNAUTH, FEDERMAN, & LAMBERT
relation expected from atomic physics. The apparent
disagreement in the 7Li/6Li isotope ratio for the second
component, between our current and previous measurements, could be due to an undetected CCD defect and/or
the sole use of the observational errors in determining the
uncertainties of the profile synthesis.
In interpreting the low isotope ratio, possible scenarios
include the following: (1) an interloper providing 7Li at the
wavelength of the 6Li line, (2) fractionation of the Li isotopes leading to an incorrect isotopic ratio for Li atoms in
the gas, or (3) a local change in the isotopic ratio arising
from synthesis of Li isotopes.
Thinner clouds than detectable from our 4044 Å spectra
are revealed by spectra of the 7699 Å K i and Na D lines. If
the absorption attributed to 6Li is in fact 7Li with a column
density of about 2 108 cm2, a typical diffuse cloud should
have a K i column density of about 4 1010 cm2 and a
detectable line at 7699 Å. Its 4044 Å counterpart would have
an W < 0:1 mÅ and be indistinguishable from the noise in
the spectrum shown in Figure 1 (top). The fifth cloud shown
by Welty & Hobbs has a column density of 3:9 1010 cm2,
but it is displaced to the red from the parent 7Li line by
about 4.7 km s1, not the by 7.1 km s1 that is the isotopic
shift. However, inclusion of this red component into the 7Li
profile synthesis does not resolve the low isotope ratio.
Even thinner clouds are detectable in the Na D lines.
Welty & Hobbs (2001) found two additional clouds to the
red of the two in Figure 1. The velocities are +10.6 and
+14.0 km s1 with Na i column densities of 4:9 1010 cm2
and 2:5 1010 cm 2, respectively (D. Welty 2001, private
communication). The 10.6 km s1 cloud’s stronger 7Li line
needs to be at +11.1 km s1 to be mistaken for 6Li in the
weaker of our two clouds. However, if the +10.6 km s1
cloud has a normal Na i/Li i ratio, the expected Li i column
density is almost 2 orders of magnitude less than that
required to account for the 6Li. Incorporation of these two
additional velocity components into the profile synthesis
yields a reasonable fit to the data and obtains 7Li/6Li ratios
that are all consistent with the solar system value. Unfortunately, this scenario solves the 6Li problem by replacing it
with another one—how to account for the very discrepant
Na i/Li i and K i/Li i ratios yet normal Na i/K i ratios.
Fractionation of the Li isotopes is an improbable scenario on several accounts. Even if no 6Li is removed from
the gas, the total Li abundance must drop by a factor of
about 4, which would place the cloud off the tight relation in
Figure 7, unless K is also removed. In addition, we have
been unable to identify a plausible fractionation process:
the predicted abundances of Li-containing molecules (Kirby
& Dalgarno 1978) are orders of magnitude too low; ionization and recombination are isotopically insensitive; gas-dust
grain processes would seem not to be dependent on the mass
difference between the isotopes.5 Therefore, it is unlikely
that 7Li is hiding in another neutral species.
A close correspondence between the exceptional 7Li/6Li
ratio and that resulting from either spallation of C, N, and
O or + fusion reactions encourages speculation that the
ratio reflects freshly synthesized lithium. The speculation
faces at least two hurdles. (1) Why is the K/Li ratio of the
exceptional cloud that of a normal cloud where local synthe5 If the grains originate in circumstellar atmospheres of red giants, they
will be Li-poor in the main. Then, evaporation of grains will produce a local
dilution of the Li/K ratio but no change in the 7Li/6Li ratio.
Vol. 586
sis of Li is not invoked? (2) What is the energy source for the
particles causing the spallation and fusion reactions?
Omicron Per is a member of the Per OB2 association and
the line of sight lies close to IC 348, an active region of star
formation. The association, as would be expected, has
experienced supernovae. X Per is a O9.5 Ve star orbiting a
neutron star (Delgado-Martı́ et al. 2001), and Per is a runaway star (Hoogerwerf, de Bruijne, & de Zeeuw 2001). Consideration of the chemistry of OH and HD led Federman et
al. (1996) to the conclusion that the interstellar clouds
toward o Per are permeated by an order of magnitude
higher cosmic-ray flux than clouds seen toward other stars
in Per OB2. The enhanced cosmic rays could indicate that
we are looking at a young superbubble where dilution has
not yet occurred.
The energetics required to synthesize the observed
amount of Li toward o Per may reveal vital information for
understanding Li production in interstellar space. Assuming
spherical geometry and using Ntot(H) and n from Table 6,
we find that 2.2 solar masses of hydrogen are present in the
interstellar clouds toward o Per. Assuming that the clouds
have the solar system Li abundance (Li/H
¼ 20:5 1010 )
implies that there are 5:4 1048 atoms of Li in the clouds.
Approximately 3 ergs of energy are needed to create a single
atom of Li (Ramaty et al. 1996). Therefore, to reproduce
the entire Li content in the clouds toward o Per requires
approximately 1% of the energy output of a typical supernova explosion. Models of superbubbles (Parizot & Drury
1999) show that the Li production per SN II is approximately 1050 atoms. The superbubble picture may apply for
the interstellar clouds toward o Per. We note that Li is not
produced solely through supernova explosions.
Studies show large variable X-ray fluxes in IC 348
(Preibisch, Zinnecker, & Herbig 1996) that are attributed to
flares produced in T Tauri stars (Preibisch, Neuhäuser, &
Alcalá 1995). The total X-ray luminosity is on the order of
1032 ergs s1 (Preibisch et al. 1996). If all 116 X-ray sources
(Preibisch et al. 1996) contribute, with flares lasting on the
order of 10 hr, the total energy produced is 4 1038 ergs,
enough for about 1038 atoms of 6Li. This calculation
assumes that all flare energy goes into the creation of 6Li.
The amount of 6Li that can be created by stellar flares is
about 10 orders of magnitude less than that needed to reproduce the entire 6Li content of the interstellar clouds toward
o Per. X-ray flares with energies on the order of 1038 ergs
have been reported (Preibisch et al. 1995). Even in these
extreme cases, 6Li production is negligible; therefore X-ray
flares cannot produce the observed 6Li abundance. A further complication is that this process is not isotope selective
since 7Li is formed as well.
8. CONSTRAINTS ON STELLAR SOURCE OF 7Li
The isotope ratios given in Table 5 provide important
constraints on stellar production pathways for 7Li (Reeves
1993; Steigman 1993). The abundance of 7Li arises from several processes: BBN, GCR spallation reactions, and stellar
production mechanisms. During the early universe, BBN
supplied about 10% of the present 7Li and a negligible
amount of 6Li. GCR spallation reactions contribute
another 10%–25% of 7Li and account for the present
abundance of 6Li. Although the nature of the stellar source
of 7Li is still unclear, it plays the most important role in Li
production today.
INTERSTELLAR 7Li/6Li RATIOS
No. 1, 2003
There has been much debate over which stellar source
dominates 7Li production throughout the lifetime of the
Galaxy. As noted in x 1, the various stellar sources of 7Li are
thought to be primarily AGB and RGB stars, SN IIs, and
possibly novae. Although AGBs have been observed with
the largest Li abundances, recent models of Galactic chemical evolution have all but eliminated AGB stars as contributors to the Galactic 7Li abundance (Charbonnel &
Balachandran 2000; Romano et al. 2001). As for the process in SN IIs, the yields of 11B and therefore 7Li have
been called into question. Alibes, Labay, & Canal (2001)
suggest a reduction by a factor of 2 in the yields of Woosley
& Weaver (1995). A problem with the source being novae is
the lack of observational evidence (Andrea, Drechsel, &
Starrfield 1994; Matteucci, D’Antona, & Timmes 1995).
Since stars have the greatest effect on the 7Li production
today, constraints are needed to isolate the stellar source(s).
Knowledge of the 6Li abundance yields the amount of 7Li
produced via GCR spallation because the isotope ratio predicted from these reactions (R76, GCR) is known with sufficient accuracy (Ramaty et al. 1997; Lemoine et al. 1998).
The isotope ratio used here is R76;GCR ¼ 1:6, which is midway between the values presented by Ramaty et al. (1997).
If the 7Li abundance in halo stars represents the primordial
abundance (Ryan et al. 2000, 2001; Suzuki et al. 2000), then
the BBN contribution is fairly well known. Additional support for Li in halo stars representing the primordial abundance comes from recent determinations of the 7Li/6Li ratio
(Smith, Lambert, & Nissen 1993; Hobbs & Thorburn 1994,
1997; Hobbs, Thorburn, & Rebull 1999). These measurements show that there has been little or no nuclear depletion
of 7Li during the lifetimes (12 Gyr) of these stars. Observations of 6Li in halo stars also yield a measure of the contamination of the primordial 7Li through GCR spallation
reactions. Recent theoretical determinations of the primordial Li abundance find AðLiÞp ¼ 2:09 (Suzuki et al. 2000), or
ðLi=HÞp ¼ 1:23 1010 . The following equation allows the
removal of the ‘‘ known ’’ sources of 7Li from the present
interstellar abundance. 7Listars represents the combined contribution from all stellar sources (AGBs, RGBs, SN IIs, and
novae),
7
Listars ¼ LiISM 6 LiISM R76;GCR 7 Lip :
ð5Þ
Also included in 7Listars is the amount of destruction by stellar astration.
The Li abundances corrected for our estimates of depletion onto grains and the amount of 7Li produced through
GCR spallation reactions are presented in Table 11. The
amount of 7Li produced by all stellar sources (using our
depletion index) is similar, approximately 7 Li=H TABLE 11
Constraints on the Stellar 7Li Source
(7Li/H)ISM
(7Li/H)GCR
(7Li/H)stars
(1.8 0.8) 109
(1.6 0.8) 109
(1.8 0.9) 109
(1.9 0.8) 109
(1.8 0.5) 109
(4.8 3.3) 1010
6.7 1010
(3.8 3.5) 1010
2.6 1010
3.7 1010
(1.2 0.9) 109
(8.4 4.8) 1010
(1.3 1.4) 109
(1.5 0.8) 109
(1.3 0.4) 109
Note.—Li/H abundance corrected for depletion using our
depletion index.
283
1 109 for all lines of sight. A larger spread arises
(5:7 1012 to 1:5 109 ) when utilizing a constant
depletion index (Welty & Hobbs 2001), suggesting that use
of a constant depletion index may not be appropriate. Since
7Li production through the various stellar sources in our
sample of sight lines is essentially constant, we infer that one
or two stellar processes dominate Li production in the solar
neighborhood. This may not be unreasonable given the
relatively small volume of the Galaxy studied.
Comparison with other elements and isotopic ratios are
needed as well. A recent study (Sofia, Meyer, & Cardelli
1999) showed that several lines of sight toward Orion
exhibit large interstellar Sn (an s-process element) abundances. Sofia et al. (1999) attribute this Sn overabundance to
AGB stars. Since AGB stars and SN II explosions also generate Li, simultaneous observations of Li, s-process, and
r-process elements are needed to constrain the amounts of
Li production by these two stellar sources. Comparisons of
r-process and s-process elements in stars are important for
determining which stellar processes contribute to the production of these elements. Knowledge of the 12C/13C and
oxygen isotope ratios is useful in constraining the production of Li in RGB stars. For instance, there could be a low
12C/13C ratio in stars with high Li abundance (Sackmann &
Boothroyd 1999), but Charbonnel & Balachandran (2000)
predict that the high 7Li abundances will be followed by the
rapid destruction of both 7Li and 12C. Therefore, stellar and
interstellar studies are required to disentangle the contributions from the different stellar sources.
9. CONCLUSIONS
The interstellar Li abundance and the 7Li/6Li isotope
ratio were measured toward several bright stars in the solar
neighborhood. These observations form the highest resolution (R 360; 000) survey of interstellar Li to date. From
these data it appears that the Li abundance has not significantly changed since the formation of the solar system and
that the solar system value of 12 for the 7Li/6Li ratio represents most of the gas in the solar neighborhood. One
exception is the LOS toward o Per. This sight line exhibits a
low 7Li/6Li ratio of about 2, the value expected from models
of GCR spallation reactions, which is consistent with our
previous determination (Knauth et al. 2000) from lower resolution observations.
If a value of about 12 was the starting condition for the
gas toward o Per, the decrease in the ratio to its present
value requires an increase in the Li abundance by an order
of magnitude. However, an enhancement in the total Li
abundance is not seen. The elemental K/Li ratio for this gas
is not unusual either, although K and Li are produced via
different nucleosynthetic pathways. The most likely scenario
for the ISM toward o Per is that the initial Li was enhanced
through GCR spallation reactions occurring as a result of
recent supernovae grouped in a superbubble. The cloud
containing a low isotope ratio contains about 20% of the
total Li content of the clouds. Therefore, determination of a
total Li abundance may mask the true Li abundance for
each interstellar cloud. Although the supernova scenario
explains many of the observations, this hypothesis does not
address the similarity of the K/Li ratio with results for other
sight lines. Fractionation cannot account for the low
7Li/6Li ratio because the abundance of LiH is too small.
There is also evidence for enhanced X-ray activity in the
284
KNAUTH, FEDERMAN, & LAMBERT
cluster IC 348 caused by stellar flares in T Tauri stars
(Preibisch et al. 1996). While 6Li can be produced during
stellar flares (e.g., Deliyannis & Malaney 1995; Ramaty et
al. 2000b), the enhanced flaring activity in OB associations
was found to play a negligible role. More typical 7Li/6Li
ratios can be obtained through the inclusion of extra velocity components that are detected in Na i absorption at 5895
Å but not seen in K i absorption at 7699 Å, although
this comes at the expense of highly disparate Na i/Li i and
K i/Li i ratios. Therefore, the exceptional isotopic ratio
toward o Per remains a puzzle. Further observations of
other stars in or near IC 348 are needed to probe the extent
of the region containing a low 7Li/6Li ratio.
Observations were made of X Per, a nearby X-ray binary.
It is believed that X Per and other X-ray binary systems are
formed after a supernova explosion (e.g., Delgado-Martı́ et
al. 2001). The direction toward X Per shows almost an order
of magnitude smaller Li and K abundances than toward
other sight lines. These smaller abundances suggest
enhanced depletion along the LOS. Unfortunately, the S/N
of the X Per spectrum was too low to allow a useful probe of
the region containing the low isotope ratio.
The stars in Sco OB2 were studied since they closely
matched the situation present toward o Per. These stars are
in an OB association where star formation is occurring in
the Oph molecular cloud. There is evidence of a past
supernova because Oph is a runaway star (Hoogerwerf et
al. 2001). Why is there no evidence for enhanced 6Li production in the presence of active star formation ( Oph molecular cloud) and past SN II explosions? Is the LOS toward o
Per really that rare? More observations are necessary to
determine whether we are experiencing a selection effect; to
date, only three sight lines in each OB association have been
studied with sufficient precision.
An additional result of our study involves the LOS
toward 20 Aql, which resides far from active star-forming
regions. The 7Li/6Li ratio is similar to the solar system
value, but the K/Li ratio is almost double the expected
value. The K/H abundance is similar to the abundance
found for other sight lines, indicating that there is less Li.
The LOS toward 20 Aql lies on the edge of Radio Loop I
(Hayakawa et al. 1979; Sembach, Savage, Tripp 1997), a
superbubble. However, the timescale needed to destroy Li
through proton burning is significantly longer than the age
of bubble. Another explanation is required to account for
the low Li/H abundance. The low Li abundance toward 20
Aql could be evidence for Li dilution in an old superbubble
(Parizot & Drury 1999; Knauth et al. 2000). Observations of
Vol. 586
Li i and K i toward other stars in the vicinity of 20 Aql are
necessary to gain further insight.
We place additional interstellar constraints on the 7Li
produced by a stellar source. Our results indicate that there
is essentially a constant production of 7Li in stars in the
solar neighborhood, suggesting one or two processes dominate. Observations of interstellar and stellar 7Li, s-process,
and r-process elements can be of great value in constraining
the Li production by the various stellar sources.
For future studies of interstellar Li, knowledge of the K i
abundance in conjunction with the molecular content can
be used to find sight lines that contain observable amounts
of Li. For instance, the results of this survey suggest other
targets in Sco OB2. These stars, which are fainter than
X Per, can be observed at lower resolving power
(R 180; 000) because their interstellar spectra reveal simple velocity structure. Preliminary results for these stars
indicate that the 7Li/6Li ratios are similar to the solar system value. Thus, lines of sight with low 7Li/6Li ratios
appear to be rare. Observations of stars in the vicinity of IC
348 (in Per OB2) are planned to help clarify the picture for
gas toward o Per. In addition, data from the Hubble Space
Telescope on the interstellar 11B/10B ratio toward stars in
Per OB2 are being analyzed. The interstellar abundance of
9Be and the 11B/10B ratio (Hebrard et al. 1997; Lambert et
al. 1998) are crucial to our understanding of Li evolution.
Knowledge of 9Be and 10B further constrain the amount of
GCR spallation occurring along the line of sight, since
GCR spallation is the only production route for these light
elements. Observations of interstellar 9Be are not yet feasible because its gas-phase abundance is extremely low
(9Be=H 7 1013 ; Hebrard et al. 1997). The amount of
11B also yields constraints on SN II explosions because
about 50% of 11B is thought to arise from neutrino-induced
spallation reactions (Woosley et al. 1990; Woosley &
Weaver 1995). Only through a more complete study of the
abundances for a variety of atomic species will the lightelement puzzle be resolved.
The McDonald Observatory data presented here formed
the basis for David Knauth’s Ph.D. dissertation. It is a
pleasure to thank the excellent staff at McDonald Observatory, especially David Doss, for assistance with the instrumental setup. We also thank János Zsargó for the use of his
code. This research made use of the Simbad database,
operated at CDS, Strasbourg, France. This research was
supported in part by NASA LTSA grant NAG 5-4957.
REFERENCES
Cassé, M., Lehoucq, R., & Vangioni-Flam, E. 1995, Nature, 373, 318
Allende Prieto, C., Lambert, D. L., & Asplund, M. 2001, ApJ, 556, L63
———. 2002, ApJ, 573, L137
Charbonnel, C., & Balachandran, S. C. 2000, A&A, 359, 563
Alibes, A., Labay, J., & Canal, R. 2001, A&A, 370, 1103
Code, A. D., Bless, R. C., Davis, J., & Brown, H. R. 1976, ApJ, 203, 417
Anders, E., & Grevesse, N. 1989, Geochim. Cosmochim. Acta, 53, 197
Crawford, I. A., & Barlow, M. J. 1996, MNRAS, 280, 863
Andrea, J., Drechsel, H., & Starrfield, S. 1994, A&A, 291, 869
Crawford, I. A., Barlow, M. J., Diego, F., & Spyromilio, J. 1994, MNRAS,
Bevington, P. R., & Robinson, D. K. 1992, Data Reduction and Error
266, 903
Analysis for the Physical Sciences (Boston: WCB McGraw-Hill), 156
Delgado-Martı́, H., Levine, A. M., Pfahl, E., & Rappaport, S. A. 2001,
Boesgaard, A. M., Deliyannis, C. P., King, J. R., & Stephens, A. 2001, ApJ,
ApJ, 546, 455
Deliyannis, C. P., & Malaney, R. A. 1995, ApJ, 453, 810
553, 754
Bohlin, R. C., Savage, B. D., & Drake, J. F. 1978, ApJ, 224, 132
Diplas, A., & Savage, B. D. 1994, ApJS, 93, 211
Boothroyd, A. I., Sackmann, I.-J., & Wasserburg, G. J. 1994, ApJ, 430,
Draine, B. T. 1978, ApJS, 36, 595
L77
Edvardsson, B., Andersen, J., Gustafsson, B., Lambert, D. L., Nissen,
———. 1995, ApJ, 442, L21
P. E., & Tomkin, J. 1993, A&A, 275, 101
Cameron, A. G. W., & Fowler, W. A. 1971, ApJ, 164, 111
Federman, S. R., Strom, C. J., & Good, J. C. 1991, AJ, 102, 1393
Federman, S. R., Strom, C. J., Lambert, D. L., Cardelli, J. A., Smith, V. V.,
Cardelli, J. A., & Ebbets, D. C. 1994, in Calibrating Hubble Space Tele& Joseph, C. L. 1994, ApJ, 424, 772
scope, ed. J. C. Blades & S. J. Osmer (Baltimore: STScI), 322
Cardelli, J. A., Mathis, J. S., Ebbets, D. C., & Savage, B. D. 1993, ApJ, 402,
Federman, S. R., Weber, J., & Lambert, D. L. 1996, ApJ, 463, 181
L17
Ferlet, R., & Dennefeld, M. 1984, A&A, 138, 303
No. 1, 2003
INTERSTELLAR 7Li/6Li RATIOS
Fields, B. D., Olive, K. A., Cassé, M., & Vangioni-Flam, E. 2001, A&A,
370, 634
Hanson, M. M., Snow, T. P., & Black, J. H. 1992, ApJ, 392, 571
Hayakawa, S., Kato, T., Nagase, F., Yamashita, K., & Tanaka, Y. 1979,
PASJ, 31, 71
Hebrard, G., Lemoine, M., Ferlet, R., & Vidal-Madjar, A. 1997, A&A,
324, 1145
Higdon, J. C., Lingenfelter, R. E., & Ramaty, R. 1998, ApJ, 509, L33
Hobbs, L. M. 1974a, ApJ, 188, L67
———. 1974b, ApJ, 191, 381
———. 1984, ApJ, 286, 252
Hobbs, L. M., & Thorburn, J. A. 1994, ApJ, 428, L25
———. 1997, ApJ, 491, 772
Hobbs, L. M., Thorburn, J. A., & Rebull, L. M. 1999, ApJ, 523, 797
Hoogerwerf, R., de Bruijne, J. H. J., & de Zeeuw, P. T. 2001, A&A, 365, 49
Howarth, I. D., Price, R. J., Crawford, I. A., & Hawkins, I. 2002, MNRAS,
335, 267
Hudson, R. D., & Carter, V. L. 1965a, Phys. Rev., 137, A1648
———. 1965b, Phys. Rev., 139, A1426
———. 1967a, J. Opt. Soc. Am., 57, 651
———. 1967b, J. Opt. Soc. Am., 57, 1471
Kirby, K., & Dalgarno, A. 1978, ApJ, 224, 444
Knauth, D. C., Federman, S. R., Lambert, D. L., & Crane, P. 2000, Nature,
405, 656
Knauth, D. C., Federman, S. R., Pan, K., Yan, M., & Lambert, D. L. 2001,
ApJS, 135, 201
Lambert, D. L., Sheffer, Y., & Crane, P. 1990, ApJ, 359, L19
Lambert, D. L., Sheffer, Y., & Federman, S. R. 1995, ApJ, 438, 740
Lambert, D. L., Sheffer, Y., Federman, S. R., Sofia, U. J., & Knauth, D. C.
1998, ApJ, 494, 614
Lemoine, M. 1995, in The Light Element Abundances, ed. P. Crane
(Garching: ESO), 350
Lemoine, M., Ferlet, R., Vidal-Madjar, A., Emerich, C., & Bertin, P. 1993,
A&A, 269, 469
———. 1995, A&A, 298, 879
Lemoine, M., Vangioni-Flam, E., & Cassé, M. 1998, ApJ, 499, 735
Lingenfelter, R. E., Ramaty, R., & Kozlovsky, B. 1998, ApJ, 500, L153
Lupton, R. 1993, Statistics in Theory and Practice (Princeton: Princeton
Univ. Press), 100
Marr, G. V., & Creek, D. M. 1968, Proc. R. Soc. London A, 304, 233
Mason, K. O., White, R. E., Sanford, P. W., Hawkins, F. J., Drake, J. F., &
York, D. G. 1976, MNRAS, 176, 193
Matteucci, F., D’Antona, F., & Timmes, F. X. 1995, A&A, 303, 460
Meneguzzi, M., Audouze, J., & Reeves, H. 1971, A&A, 15, 337 (MAR)
Meyer, D. M., Hawkins, I., & Wright, E. L. 1993, ApJ, 409, L61
Morton, D. C. 1991, ApJS, 77, 119
Papaj, J., Krelowski, J., & Wegner, W. 1991, MNRAS, 252, 403
Parizot, E., & Drury, L. 1999, A&A, 349, 673
Pequignot, D., & Aldrovandi, S. M. V. 1986, A&A, 161, 169
Perryman, M. A. C., et al. 1997, A&A, 323, L49
Plez, B., Smith, V. V., & Lambert, D. L. 1993, ApJ, 418, 812
Preibisch, T., Neuhaeuser, R., & Alcala, J. M. 1995, A&A, 304, L13
Preibisch, T., Zinnecker, H., & Herbig, G. H. 1996, A&A, 310, 456
Ramaty, R., Kozlovsky, B., & Lingenfelter, R. E. 1996, ApJ, 456, 525
Ramaty, R., Kozlovsky, B., Lingenfelter, R. E., & Reeves, H. 1997, ApJ,
488, 730
Ramaty, R., & Lingenfelter, R. E. 1999, Ap&SS, 265, 71
Ramaty, R., Scully, S. T., Lingenfelter, R. E., & Kozlovsky, B. 2000a, ApJ,
534, 747
285
Ramaty, R., Tatischeff, V., Thibaud, J. P., Kozlovsky, B., & Mandzhavidze, N. 2000b, ApJ, 534, 207
Reddy, B. E., Tomkin, J., Lambert, D. L., & Allende Prieto, C. 2003,
MNRAS, in press
Reeves, H. 1974, ARA&A, 12, 437
———. 1993, A&A, 269, 166
Reeves, H., Audouze, J., Fowler, W. A., & Schramm, D. 1973, ApJ, 179,
909
Reeves, H., Fowler, W. A., & Hoyle, F. 1970, Nature, 226, 727
Romano, D., Matteucci, F., Ventura, P., & D’Antona, F. 2001, A&A, 374,
646
Ryan, S. G., Beers, T. C., Olive, K. A., Fields, B. D., & Norris, J. E. 2000,
ApJ, 530, L57
Ryan, S. G., Kajino, T., Beers, T. C., Suzuki, T. K., Romano, D.,
Matteucci, F., & Rosolankova, K. 2001, ApJ, 549, 55
Sackmann, I.-J., & Boothroyd, A. I. 1999, ApJ, 510, 217
Sandner, W., Gallagher, T. F., Safinya, K. A., & Gounand, F. 1981, Phys.
Rev. A, 23, 2732
Sansonetti, C. J., Richou, B., Engelman, Jr. R., & Radziemski, L. J. 1995,
Phys. Rev. A, 52, 2682
Savage, B. D., Bohlin, R. C., Drake, J. F., & Budich, W. D. 1977, ApJ, 216,
291
Sembach, K. R., Savage, B. D., & Tripp, T. M. 1997, ApJ, 480, 216
Smith, V. V., & Lambert, D. L. 1989, ApJ, 345, L75
———. 1990, ApJ, 361, L69
Smith, V. V., Lambert, D. L., & Nissen, P. E. 1993, ApJ, 408, 262
Smith, V. V., Plez, B., Lambert, D. L., & Lubowich, D. A. 1995, ApJ, 441,
735
Snell, R. L., & Vanden Bout, P. A. 1981, ApJ, 250, 160
Snow, T. P., Black, J. H., van Dishoeck, E. F., Burks, G., Crutcher, R. M.,
Lutz, B. L., Hanson, M. M., & Shuping, R. L. 1996, ApJ, 465, 245
Sofia, U. J., Cardelli, J. A., Guerin, K. P., & Meyer, D. M. 1997, ApJ, 482,
L105
Sofia, U. J., Fitzpatrick, E. L., & Meyer, D. M. 1998, ApJ, 504, L47
Sofia, U. J., Meyer, D. M., & Cardelli, J. A. 1999, ApJ, 522, L137
Starrfield, S., Truran, J. W., Sparks, W. M., & Arnould, M. 1978, ApJ, 222,
600
Steigman, G. 1993, ApJ, 413, L73
———. 1996, ApJ, 457, 737
Suzuki, T. K., Yoshii, Y., & Beers, T. C. 2000, ApJ, 540, 99
Traub, W. A., & Carleton, N. P. 1973, ApJ, 184, L11
Vanden Bout, P. A., Snell, R. L., Vogt, S. S., & Tull, R. G. 1978, ApJ, 221,
598
Vangioni-Flam, E., Cassé, M., & Audouze, J. 2000, Phys. Rep., 333
Vangioni-Flam, E., Ramaty, R., Olive, K. A., & Cassé, M. 1998, A&A,
337, 714
Wannier, P., Andersson, B.-G., Penprase, B. E., & Federman, S. R. 1999,
ApJ, 510, 291
Wasserburg, G. J., Boothroyd, A. I., & Sackmann, I.-J. 1995, ApJ, 447,
L37
Welty, D. E., & Hobbs, L. M. 2001, ApJS, 133, 345
Welty, D. E., Hobbs, L. M., & Kulkarni, V. P. 1994, ApJ, 436, 152
White, R. E. 1986, ApJ, 307, 777
Wielen, R., Fuchs, B., & Dettbarn, C. 1996, A&A, 314, 438
Wielen, R., & Wilson, T. L. 1997, A&A, 326, 139
Woosley, S. E., Hartmann, D. H., Hoffman, R. D., & Haxton, W. C. 1990,
ApJ, 356, 272
Woosley, S. E., & Weaver, T. A. 1995, ApJS, 101, 181
Zsargó, J., & Federman, S. R. 2002, ApJ, submitted
The Astrophysical Journal, 594:664, 2003 September 1
# 2003. The American Astronomical Society. All rights reserved. Printed in U.S.A.
ERRATUM: ‘‘ AN ULTRA–HIGH-RESOLUTION SURVEY OF THE INTERSTELLAR 7Li/6Li ISOTOPE
RATIO IN THE SOLAR NEIGHBORHOOD ’’ (ApJ, 586, 268 [2003])
David C. Knauth, S. R. Federman, and David L. Lambert
An error occurred in Table 10 because author corrections were not transferred faithfully during the
production process. The corrected table appears below. The Press sincerely regrets this error.
TABLE 10
Kinetic Temperature and Turbulent Velocity
Star
o Per.........................
X Perd ......................
Ophd .....................
Oph .......................
20 Aql ......................
Tka
(K)
vturb b
(km s 1)
vturb c
(km s 1)
98 42
900d
520
788
220
61 27
63 27
287 124
770 330
0.16 0.10
0.70d
0.77
0.43
0.37
0.37 0.23
0.30 0.18
0.47 0.29
0.39 0.24
0.15 0.06
1.20d
1.05
1.00
0.53
0.30 0.13
0.21 0.09
0.67 0.29
0.97 0.42
a
Tk derived from both b-values.
Uses calculated Tk.
c Assumes T = 100 K.
k
d Extreme values of the b-values ( 1 ) were used to arrive at a
solution for Tk and vturb ; no error bars are given.
b
664